=Paper=
{{Paper
|id=Vol-3180/paper-156
|storemode=property
|title=Overview of SnakeCLEF 2022: Automated Snake Species Identification on a Global Scale
|pdfUrl=https://ceur-ws.org/Vol-3180/paper-156.pdf
|volume=Vol-3180
|authors=Lukáš Picek,Marek Hruz,Andrew M. Durso,Isabelle Bolon
|dblpUrl=https://dblp.org/rec/conf/clef/PicekHDB22
}}
==Overview of SnakeCLEF 2022: Automated Snake Species Identification on a Global Scale==
https://ceur-ws.org/Vol-3180/paper-156.pdf
Overview of SnakeCLEF 2022: Automated Snake
Species Identification on a Global Scale
Lukáš Picek1 , Marek Hrúz1 , Andrew M. Durso2 and Isabelle Bolon4
1
Department of Cybernetics, Faculty of Applied Sciences, University of West Bohemia, Czechia
2
Department of Biological Sciences, Florida Gulf Coast University, Florida, USA
3
Institute of Global Health, Department of Community Health and Medicine, University of Geneva, Switzerland
Abstract
The main goal of the third year of the SnakeCLEF challenge was to provide an evaluation platform that
helps track the performance of AI-driven methods for snake species recognition systems on a global
scale and allows direct comparison with human experts. We ran two challenges separately for humans —
experts and novices — and AI methods in order to lay the groundwork for future comparison between
human and machine-based snake species identification. We have provided 187,129 snake observations
with 318,532 photographs — 270,251 for training and 48,281 for testing — of 1,572 snake species collected
in 208 countries. The human performance evaluation was conducted on a tailored subset with 150
images derived from the full test set. We report (i) a description of the provided data, (ii) evaluation
methodology and principles, (iii) an overview of the methods submitted by the participating teams, and
(iv) a discussion of the obtained results.
Keywords
LifeCLEF, SnakeCLEF, fine grained visual categorization, global health, epidemiology, snake bite, snake,
reptile, benchmark, biodiversity, species identification, machine learning, computer vision, classification
1. Introduction
A robust image-based identification system for snake species is an important goal for biodiversity,
conservation, and global health. With over half a million victims of death and disability from
venomous snakebite annually, such a system could significantly improve eco-epidemiological
data and treatment outcomes (e.g. selection of specific antivenoms) [1, 2]. Importantly, most
herpetological expertise and most snake images are concentrated in developed countries where
snake diversity is relatively low and snakebite is not a major public health concern. In contrast,
remote parts of developing countries tend to lack expertise and images, even in areas where
snake diversity is high and snakebites are common [3, 4]. Thus, snake species identification
assistance has a bigger potential to save lives in areas with the least information.
A primary difficulty of snake species identification lies in the high intra-class and low inter-
class variance in appearance, which may depend on geographic location, color morph, sex, or
age. At the same time, many species are visually similar to other species, i.e., mimicry (Figure 1).
CLEF 2022: Conference and Labs of the Evaluation Forum, September 5–8, 2022, Bologna, Italy
$ picekl@kky.zcu.cz (L. Picek)
0000-0002-6041-9722 (L. Picek); 0000-0002-2287-0985 (M. Hrúz); 0000-0002-3008-7763 (A. M. Durso);
0000-0001-5940-2731 (I. Bolon)
© 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073 CEUR Workshop Proceedings (CEUR-WS.org)
Furthermore, our knowledge of which snake species occur in which countries is incomplete,
and it is common that most or all images of a given snake species might originate from a small
handful of countries or even a single country. Furthermore, many snake species resemble species
found on other continents, with which they are entirely allopatric. Incorporating metadata on
the geographic origin of an unidentified snake almost always narrows down the possible correct
identifications considerably because only about 125 of the approximately 3,900 snake species
co-occur in any given location [5]. It is known that more widespread species with more images
are over-predicted relative to rare species with few images [6], and this can be a particularly
vexing problem when trying to predict the identity of species that are widespread across areas
of the world with few images.
The main goal of the SnakeCLEF 2022 competition was to provide a reliable evaluation ground
for automatic snake species recognition. Like other LifeCLEF competitions, the SnakeCLEF 2022
competition was hosted on Kaggle1 primarily to attract machine learning experts to participate
and present their ideas.
Figure 1: Harmless mimic species Cemophora coccinea ssp. coccinea (top row) and poisonous lookalike
species. Micrurus pyrrhocryptus, Micrurus ibiboboca, and Micrurus nigrocinctus (left to right, bot. row).
©roadmom–iNaturalist, ©Anthony Damiani–iNaturalist, ©Adam Cushen–iNaturalist, ©Alexander Guiñazu–
iNaturalist, ©Tarik Câmara–iNaturalist, and ©Cristhian Banegas–iNaturalist.
1
https://www.kaggle.com/competitions/fungiclef2022
Figure 2: Two snake observations from SnakeCLEF2022 dataset — three images for each individual.
©André Giraldi – iNaturalist, ©Harshad Sharma – iNaturalist.
Table 1
Details of the SnakeCLEF 2022 datasets and their comparison with previous editions.
Dataset Species Images Observation Countries min / max samples
SnakeCLEF 2020 783 259,214 × 145 19 14,433
SnakeCLEF 2021 772 386,006 × 188 10 22,163
SnakeCLEF 2022 1,572 318,532 187,129 208 5 6,472
SnakeCLEF 2022–Training 1,572 270,251 158,698 207 3 5,518
SnakeCLEF 2022–Test 1,572 48,281 28,431 183 2 954
1.1. Dataset
For this year, the dataset used in previous editions [7, 8] has been extended with new and
rare species. The number of species was doubled and the number of images from remote
geographic areas with no or just a few samples was increased considerably, i.e., the uneven
species distributions across all the countries was straightened. The SnakeCLEF 2022 dataset
is based on 187,129 snake observations — including some instances of multiple images of the
same individual (refer to Figure 2) — with 318,532 photographs belonging to 1,572 snake species
and observed in 208 countries. The dataset has a heavy long-tailed class distribution, where
the most frequent species (Natrix natrix) is represented by 6,472 images and the least frequent
species just by 5 samples. The difference in the number of images between the species with the
most and fewest was reduced by an order of magnitude relative to SnakeCLEF2021. All the data
were gathered from the online biodiversity platform iNaturalist2 . Additional dataset parameters
and comparison with previous editions are listed in Table 1.
2
https://www.inaturalist.com/
Figure 3: Left: Worldwide snake species distribution, i.e., The number of species found in each country.
Large countries in the tropics (Brazil, Mexico, Colombia, India, and Indonesia) have more than 300
species. Right: Percentage of snake species per country included in the SnakeCLEF 2022 dataset. The
countries with adequate species coverage are those from Europe, Oceania, and North America, i.e., the
countries with the smallest diversity.
For testing, two sets were created: (i) the full test set for a machine evaluation, with 48,280
images from 28,431 observations, and (ii) the subset from the full test set with 150 observations,
tailored for the human performance evaluation. Unlike in other LifeCLEF competitions, where
the final testing set remained undisclosed, we provided the test data without labels to the
participants. To prevent over fitting to the leaderboard, the evaluation method was composed
of two stages; the first being the public leaderboard where the user scores were calculated on
an unknown 20% of the test set, and the second a private leaderboard where participants were
scored on the remaining part of the test set. In addition to image data, we provide:
• human verified species labels that allow up-scaling to higher taxonomic ranks,
• the country-species mapping file describing country-species presence to allow better
regularization towards all geographical locations, based on The Reptile Database [9], and
• information about endemic species — species that occur only in one geographical region,
e.g., Australia or Madagascar.
Geographical information, i.e., state/province and country labels, was included for approxi-
mately 95% of the training and test images. Additionally, we provide a mapping matrix (MM𝑐𝑠 )
describing country-species presence to allow better worldwide regularization.
{︃
1 if species S ∈ country C,
MM𝑐𝑠 = (1)
0 otherwise.
Unlike last year’s dataset, where the vast majority (77%) of all images came from the United
States and Canada, the SnakeCLEF 2022 dataset includes just a fraction of the data (28.3%) from
the United States and Canada. The rest of the data is distributed across remaining regions,
e.g., Europe, Asia, Africa, Australia and Oceania. This was achieved by limiting the number of
observations per species in a given country to 400. The estimated worldwide snake distribution
and their coverage is visualized in Figure 3.
1.2. Timeline
The SnakeCLEF 2022 competition was announced together with the data in late February 2022
through the LifeCLEF, Kaggle, and FGVC challenge pages. Anyone with research ambitions was
allowed to register and participate in the competition. The test data were provided jointly with
the training data allowing continuous evaluation. Each team could submit up to 2 submissions
a day. The competition deadline was May 16, setting the competition for roughly three months.
Participants had to submit CSV files containing the Top1 prediction for each snake observation.
Once the submission phase was closed (mid-May), the participants were allowed to submit
post-competition submissions to evaluate their ablation studies.
1.3. Evaluation Protocol
The main goal of this challenge was to build a system that is capable of recognizing 1,572
snake species based on the given snake observation — unseen set of images — and relevant
geographical location. As a main metric, we use the macro F1 score (F𝑚
1 ). The F1 is defined as
𝑚
the mean of class-wise F1 scores:
𝑁
1 ∑︁ 𝑃𝑠 × 𝑅 𝑠
F𝑚
1 = 𝐹1𝑠 , 𝐹1𝑠 = 2 × , (2)
𝑁 𝑃𝑠 + 𝑅 𝑠
𝑠=0
where 𝑠 is species index, 𝑁 equals to the number of classes in a training set. The F1 score
for each class represents harmonic mean of the class precision 𝑃𝑠 and recall 𝑅𝑠 . This type of
evaluation is suitable for data with long-tail distribution, since the quantity of samples from
individual classes does not effect the outcome. For the additional evaluation of the performance
of the teams we also compute the micro classification accuracy, which is a ratio between the
number of correctly classified samples and all samples.
1.4. Working Notes
All participants were asked to provide code and a Working Note paper — a technical report with
information needed to reproduce the results of all submissions. All submitted Working Notes
were reviewed by 2–3 reviewers with a decent publication history in the field of Computer
Vision and Machine Learning, ensuring a sufficient reproducibility and quality. The review
process was single-blind and offered up to two rebuttals. The acceptance rate was 66.66%.
2. Results
The best performing team achieved F𝑚 1 of 86.47% on the private part of the test set and 94.01%
classification accuracy on the full test set. Both scores are the best ones in the given category.
We observe a steady decrease in the F𝑚 1 score for the first three teams achieving 84.87% and
82.65% F𝑚 1 score. Then there are several significant drops in performance. Similar behavior is
present in the classification accuracy with the second team achieving 93.29% and the third one
achieving 93.53%. It can be seen that the accuracy is not correlated with the F𝑚 1 score which
Accuracy [%] Macro averaged F1 [%]
0
20
40
60
80
100
0
20
40
60
80
100
GG 94.01 GG 86.47
92 90.28
Secret;Weapon 93.29 Secret;Weapon 84.87
94 89.96
SAI 93.53 SAI 82.65
94 87.55
WabuLabuDabu 90.25 WabuLabuDabu 76.47
94.67 80.23
LieDown 88.6 LieDown 72.93
91.33 78.2
comfort break 90.04 comfort break 72.01
91.33 77.92
USTC-IAT-United 89.97 USTC-IAT-United 71.97
91.33 78.26
FHDO-BCSG 88.4 FHDO-BCSG 70.8
85.33 75.43
vicii 83.54 vicii 68.56
86.67 70.87
DiamondH 81.29 DiamondH 59.5
84 65.64
BingXi 78.75 BingXi 57.02
82.67 62.13
College Slytherin 58.32 College Slytherin 44.28
54 49.04
jzsherlock 61.99 jzsherlock 42.3
60.67 45.84
evaluation for Top-25 teams is provided in Figure 4.
Got it ! 65.89 Got it ! 41.87
68 47.38
dmitrykonovalov 67.46 dmitrykonovalov 39.55
70.67 44.08
Ritu Ahmed 67.85 Ritu Ahmed 39.16
64 44.39
MAGUS_YWX 62.24 MAGUS_YWX 36.51
65.33 41.19
On the expert set, the best performing team — 4th in terms of F𝑚
YHT_MT1 46.85 YHT_MT1 32.18
52.67 40.52
SSN-CSE-Lekshmi 52.74 SSN-CSE-Lekshmi 31.92
52 36.38
asdaas1 39.17 asdaas1 26.31
45.33 33.72
Further performance evaluation for the Top-25 teams is provided in Figure 5.
7aml 47.79 7aml 22.38
53.33 26.78
Rziting 30.76 Rziting 21.04
31.33 27.41
plcrtkqrm 25.08 plcrtkqrm 17.14
28 23.33
HaHaWork 22.67 HaHaWork 15.84
26 21.62
Figure 4: Official SnakeCLEF 2022 competition results, sorted by performance on the private set.
29.19 9.9
Expert Set
Public Leaderboard
28.67 12.01
Private Leaderboard
Full Test Set
1 score — achieved impressive
suggests that some teams did not cope well with the long tail distribution. Further performance
Figure 5: Identification accuracy on the full test set and the test set for human performance evaluation.
accuracy of 94.67%. Eight teams achieved satisfactory accuracy on the expert set, around 90%.
2.1. Participants and Methods
A total of 31 teams participated in the SnakeCLEF 2022 challenge and submitted 676 submissions.
Everyone who submitted a solution better than baseline submission, i.e., random predictions,
was considered a participant. The number of participants quadrupled since last year, primarily
because Kaggle was used as an evaluation platform. More details can be found in the individual
working notes of participants [10, 11, 12, 13, 14, 15] who passed the review process, ensuring
a sufficient level of reproducibility and quality. The main outcomes we can derive from the
achieved results are as follows:
GG [10]: The winner of the challenge. They have introduced a novel architecture CoLKA-
Net based on VAN [16] (Visual Attention Network) and CoAtNet [17]. It is a combination of
large kernel attention and vision transformer. In an ablation study, the model outperforms
other tested models — ViT [18], Swin [19], VOLO [20], and ConvNeXt [21] by around 2 points
of F𝑚
1 score (78.00% to 80.10%). In addition, the team used techniques such as Label Aware
Smoothing [22], Pseudo labelling for tail classes, FixRes mitigation [23], and augmentations.
When TrivialAugment [24] was deployed during the middle stage of experimentation the team
observed a rise in F𝑚
1 of around 0.5%. Progressively, Random Erasing [25], CutMix [26] and
Mixup [27] were added which helped with regularization. The final submission score was
achieved by an ensemble of six models 2× ConvNeXt, VOLO, CoLKANet, Swin, and ViT. The
novel CoLKANet is an interesting contribution with potential outside of the scope of this
competition.
Secret;Weapon [11]: The runner-up focused solely on ViT models, ensembling three dif-
ferent variants; two Large models trained on different resolutions (384; 432) and one Huge
model (392). The observed trend from the ablation study is that larger models with higher input
resolution perform better. The ViT models were pretrained as Masked autoencoders [28] on
ImageNet-1K [29]. The team designed a new Effective Logit Adjustment Loss combining Logit
adjustment loss [30] and Class-balanced Loss [31]. They observe a better overall performance
when compared to Cross Entropy and further analyze that the improvement comes from the
tail classes. The metadata was integrated by applying an estimated species a priori distribution
in individual regions to the ViT estimates. Interestingly, when all metadata (code, endemic,
country) were added the F𝑚 1 score on validation data rose but dropped rapidly on the test data.
The best combination on the test data was when the country code was omitted. This hints at
a distribution gap between the training and test metadata.
SAI [12]: The team used MetaFormer [32]. The metadata passes through an Embedding, indi-
vidual tags (code, endemic, country) are concatenated and the final MLP is used to produce
the Meta token. The Metaformer seems suitable for integrating categorical tokens. The team
shows in an ablation study a significant raise in F𝑚1 score from 66.64% to 74.18%. The Logit
adjustment loss function is used to optimize the model. It works better than Seesaw [33] loss
but is outperformed by a post-hoc logit adjustment. Filtering the predictions based on possible
locations of species raises the F𝑚
1 score from 64.32% to 69.09%. SimCLR [34] with InfoLoss [35]
is used to train the model. The team shows significant improvement when compared to only
supervised models pre-trained on different datasets (eg. from ImageNet-1k pre-train F𝑚 1 of
63.76% to 68.83% when SimCLR is used). In accordance with the findings of previous teams,
a bigger resolution of input images leads to better scores. The final score was achieved by the
multi-scale, multi-crop, and multi-resolution model ensemble.
USTC-IAT-United [13]: The team combines models of Swin Transformer, EfficientNet, and
BEiT [36]. They experiment with several loss functions (CE Loss, Seesaw Loss, Focal Loss [37])
and show that Focal Loss performs the best, but only marginally (less than 1% F𝑚 1 score im-
provement when compared to CE). From the individual models, the EfficientNet-L2 is the best
one. Given the highest number of parameters (480M), it is expected. From this point of view,
it is interesting that Swin with 197M parameters achieves the worst score (F𝑚 1 of 61.70%) and
is outperformed by EfficientNet-B7 (F1 of 64.99%) with only 66M parameters. The authors
𝑚
experimented with a fine-grained classification technique — PIM [38] — to find informative and
discriminative features. They show the improvement on the Swin model (F𝑚 1 from 62.38% to
63.80%) but then do not use it because of the relatively big computational burden when training
the model. Finally, the models are ensembled by averaging their softmax outputs. The authors
observe a steady increase in F𝑚1 score when adding the models to the ensemble one by one. But
an unexpected drop is observed when the last model — Swin — is added. The drop is present
only in the test data and not the validation data. This opens the question of how, when, and
what models to add to the ensemble.
FHDO-BCSG [14]: The team adopted the two-stage principle of the best solution from Snake-
CLEF2021 [39]. In the first stage, the snake is detected and in the second stage, the detected
region is classified by a CNN. However, this year it was not the best choice, substantially lacking
in performance when compared to the winning team (70.7% F𝑚 1 versus 85.4% F1 ). Even so, the
𝑚
detection helps and the YOLOv5 [40] detection network improved the private F1 by 13% on aver-
𝑚
age. The team experimented with two CNN models — EfficientNet and ConvNeXt. They provide
an in-depth study of the behavior of the models when combined with different techniques of
optimizing, adding metadata, types of augmentations, and so on. The best competition score
was achieved by ensembling seven models — EfficientNet-B4 with and without object detection,
EfficientNet-v2-m with no object detection, and EfficientNet-v2-m with and without object
detection. However this score of 70.79% F𝑚 1 is only a bit ahead of a single EfficientNet-v2-m
model with score 70.23% F1 . This begs the question of how important it is to use ensemble in
𝑚
a real-life application when the improvement is limited. Lastly, the team experimented with
different types of metadata representation. In all cases, the metadata were used as a priori
distributions of snake species in different locations that were multiplied with the resulting
softmax. The worst option was to use the estimated a priori distribution, followed by the
binarized version (with a threshold of 0) and finally the post-competition best result (73.90%
F𝑚
1 ) when the binarized distribution was multiplied by the a priori country distribution of the
training dataset.
anonymous_rice [15]: The team experimented with several CNN models — ResNet, ResNext,
and EfficientNet to produce deep features of the images and concatenate them with the cate-
gorical representation of metadata. In the final solution, ResNet-101 and EfficientNet-B0 were
used. The concatenated features are inputted into the XGBoost Ensemble Classifier [41] to
produce the final classification. The nice property of the XGBoost algorithm is that the relative
importance of the ensembled features is computed. Unfortunately, the team achieved low score
of 3.6% F𝑚
1 but after the competition when the backbones were trained further, the score raised
significantly to 51.39% F𝑚1 . This shows some potential of the XGBoost algorithm that may be
interesting to study in the future.
3. Conclusions and Perspectives
This paper presents an overview and results evaluation of the third edition of the SnakeCLEF chal-
lenge organized in conjunction with the Conference and Labs of the Evaluation Forum (CLEF3 )
and LifeCLEF4 research platform [42], and FGVC9 Workshop5 — The Ninth Workshop on Fine-
Grained Visual Categorization organized within the CVPR conference. The main outcomes we
can derive from the this year’s evaluation are as follows.
Transformer-based architectures outperformed CNNs. This year various deep neural
network architectures — Convolutional Neural Networks and Transformers — were evalu-
ated; ConvNext [21], EfficientNet [43], Vision Transformer [18], Swin Transformer [19], and
MetaFormer [32]. Unlike last year, where the CNN architectures overwhelmed the performance,
Vision Transformer architectures were a vital asset for most methods submitted this year. The
second best method with F𝑚 1 score of 84.56% was based on an ensemble of exclusively ViT models
and performed slightly worse (–0.9%) than the best performing system that used a combination
of Transformer and CNN models. An ensemble of MetaFormer models achieved the third-best
score of 82.65%. It seems that Transformers and CNNs benefit from each other in an ensemble,
whereas a standalone Transformer ensemble performs better than a pure CNN ensemble which
achieved an F𝑚1 score of "only" 70.80%
Loss Function matters. Several loss functions were evaluated: Label Aware Smoothing [22],
(modified) Categorical Cross-Entropy, Seesaw [33], and Focal Loss [37]. Overall, any Loss func-
tion if used is better than standard CrossEntropy. The wining team used Label Aware Smoothing.
The runner-up used an Effective Logit Adjustment Loss and showed an improvement of around
2% of F𝑚 1 score when compared to Cross Entropy, reducing the error rate by 15%. The the third
team used Logit adjustment to outperform the Seesaw loss from an F𝑚 1 score of 76.49% to 78.57%.
The team USTC-IAT-United compared CE Loss, Seesaw Loss, and Focal Loss with EfficientNet
model. Focal Loss performed the best, but only with marginal improvement over CE Loss and
Seesaw Loss.
Self-supervision has potential. Adding unlabeled data to the train set is a welcome option
when not many observations of a species are available. The third team used the SimCLR [34]
method with InfoNCE [35] loss function to increase the F𝑚
1 score from 63.76% to 68.83% when
3
http://www.clef-initiative.eu/
4
http://www.lifeclef.org/
5
https://sites.google.com/view/fgvc9/home
compared to an ImageNet-1k pretrained models. Overall, performance on tail classes was higher
this year.
Geographical metadata improves classification performance. Most teams report ac-
curacy improvement when adding the metadata into the learning process. The second team
achieved an improvement of 10.89% in terms of the F𝑚 1 score using a simple location filter-
ing approach. The third team described an absolute improvement of 7.54% when adding the
metadata into the MetaFormer. Using the a priori country distribution from the training data
outperformed other approaches tried by the fifth team, but in real life this will frustrate users in
countries that lack many snake images to use as training data. Representation of many locations
in both training and testing data remain important and challenging.
Ensemble helps, but at what cost? Most teams used ensembling to increase the accuracy
of classification. The standard approach was to compute an average of the individual models’
decisions. Some teams used a late fusion of deep features by concatenation as an ensemble
technique. Even though the improvement in accuracy is observable (around 1 percentage point
of F𝑚
1 across the board), it would be interesting to measure the added computational complexity
vs the added accuracy. In the case of snakebite, the system’s inference time plays a crucial role.
Acknowledgments
LP was supported by the UWB grant, project No. SGS-2022-017. LP and MH were supported by
the Technology Agency of the Czech Republic, project No. SS05010008. AMD was supported
by the Florida Gulf Coast University Office of Scholarly Innovation and Student Research. The
work described herein has been supported by the Ministry of Education, Youth and Sports of
the Czech Republic, Project No. LM2018101 LINDAT/CLARIAH-CZ.
References
[1] I. Bolon, A. M. Durso, S. Botero Mesa, N. Ray, G. Alcoba, F. Chappuis, R. Ruiz de Castañeda,
Identifying the snake: First scoping review on practices of communities and healthcare
providers confronted with snakebite across the world, PLoS ONE 15 (2020) e0229989.
[2] R. Ruiz de Castañeda, A. M. Durso, N. Ray, J. L. Fernández, D. J. Williams, G. Alcoba,
F. Chappuis, M. Salathé, I. Bolon, Snakebite and snake identification: empowering neglected
communities and health-care providers with ai, The Lancet Digital Health 1 (2019) e202–
e203.
[3] A. M. Durso, R. Ruiz de Castañeda, C. Montalcini, M. R. Mondardini, J. L. Fernandez-
Marques, F. Grey, M. M. Müller, P. Uetz, B. M. Marshall, R. J. Gray, et al., Citizen science
and online data: Opportunities and challenges for snake ecology and action against
snakebite, Toxicon: X (2021) 100071.
[4] A. M. Durso, I. Bolon, A. Kleinhesselink, M. Mondardini, J. Fernandez-Marquez, F. Gutsche-
Jones, C. Gwilliams, M. Tanner, C. E. Smith, W. Wüster, et al., Crowdsourcing snake
identification with online communities of professional herpetologists and avocational
snake enthusiasts, Royal Society open science 8 (2021) 201273.
[5] U. Roll, A. Feldman, M. Novosolov, A. Allison, A. M. Bauer, R. Bernard, M. Böhm, F. Castro-
Herrera, L. Chirio, B. Collen, et al., The global distribution of tetrapods reveals a need for
targeted reptile conservation, Nature Ecology & Evolution 1 (2017) 1677–1682.
[6] A. M. Durso, G. K. Moorthy, S. P. Mohanty, I. Bolon, M. Salathé, R. Ruiz de Castañeda,
Supervised learning computer vision benchmark for snake species identification from
photographs: Implications for herpetology and global health, Frontiers in Artificial
Intelligence 4 (2021) 17.
[7] L. Picek, R. Ruiz de Castañeda, A. M. Durso, S. P. Mohanty, Overview of the snakeclef 2020:
Automatic snake species identification challenge, in: CLEF task overview 2020, CLEF:
Conference and Labs of the Evaluation Forum, Sep. 2020, Thessaloniki, Greece., 2020.
[8] L. Picek, A. M. Durso, R. Ruiz De Castañeda, I. Bolon, Overview of SnakeCLEF 2021:
Automatic snake species identification with country-level focus, in: Working Notes of
CLEF 2021 - Conference and Labs of the Evaluation Forum, 2021.
[9] P. Uetz, P. Freed, J. Hošek, et al., The reptile database, 2020. URL: https://reptile-database.
reptarium.cz/advanced_search.
[10] Y. Shen, X. Sun, Z. Zhu, When large kernel meets vision transformer: A solution for
snakeclef & fungiclef, in: Working Notes of CLEF 2022 - Conference and Labs of the
Evaluation Forum, 2022.
[11] L. Yang, X. Li, R. Song, K. Zhu, G. Li, Solution for snakeclef 2022 by tackling long-tailed
categorization, in: Working Notes of CLEF 2022 - Conference and Labs of the Evaluation
Forum, 2022.
[12] C. Zou, F. Xu, M. Wang, W. Li, Y. Cheng, Solutions for fine-grained and long-tailed snake
species recognition in snakeclef 2022, in: Working Notes of CLEF 2022 - Conference and
Labs of the Evaluation Forum, 2022.
[13] J. Yu, H. Chang, Z. Cai, G. Xie, L. Zhang, K. Lu, S. Du, Z. Wei, Z. Liu, F. Gao, F. Shuang,
An efficient model integration-based snake classification algorithm, in: Working Notes of
CLEF 2022 - Conference and Labs of the Evaluation Forum, 2022.
[14] L. Bloch, J.-F. Böckmann, B. Bracke, C. M. Friedrich, Combination of object detection,
geospatial data, and feature concatenation for snake species identification, in: Working
Notes of CLEF 2022 - Conference and Labs of the Evaluation Forum, 2022.
[15] M. Palaniappan, K. Desingu, H. Bharathi, E. A. Chodisetty, A. Bhaskar, Deep learning and
gradient boosting ensembles for classification of snake species, in: Working Notes of CLEF
2022 - Conference and Labs of the Evaluation Forum, 2022.
[16] M.-H. Guo, C.-Z. Lu, Z.-N. Liu, M.-M. Cheng, S.-M. Hu, Visual attention network, arXiv
preprint arXiv:2202.09741 (2022).
[17] Z. Dai, H. Liu, Q. V. Le, M. Tan, Coatnet: Marrying convolution and attention for all data
sizes, Advances in Neural Information Processing Systems 34 (2021) 3965–3977.
[18] A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. De-
hghani, M. Minderer, G. Heigold, S. Gelly, et al., An image is worth 16x16 words: Trans-
formers for image recognition at scale, arXiv preprint arXiv:2010.11929 (2020).
[19] Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang, S. Lin, B. Guo, Swin transformer: Hierarchical
vision transformer using shifted windows, in: Proceedings of the IEEE/CVF International
Conference on Computer Vision, 2021, pp. 10012–10022.
[20] L. Yuan, Q. Hou, Z. Jiang, J. Feng, S. Yan, Volo: Vision outlooker for visual recognition,
arXiv preprint arXiv:2106.13112 (2021).
[21] Z. Liu, H. Mao, C.-Y. Wu, C. Feichtenhofer, T. Darrell, S. Xie, A convnet for the 2020s, in:
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,
2022, pp. 11976–11986.
[22] Z. Zhong, J. Cui, S. Liu, J. Jia, Improving calibration for long-tailed recognition, in:
Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021,
pp. 16489–16498.
[23] H. Touvron, A. Vedaldi, M. Douze, H. Jégou, Fixing the train-test resolution discrepancy,
Advances in neural information processing systems 32 (2019).
[24] S. G. Müller, F. Hutter, Trivialaugment: Tuning-free yet state-of-the-art data augmentation,
in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp.
774–782.
[25] Z. Zhong, L. Zheng, G. Kang, S. Li, Y. Yang, Random erasing data augmentation, in:
Proceedings of the AAAI conference on artificial intelligence, volume 34, 2020, pp. 13001–
13008.
[26] S. Yun, D. Han, S. J. Oh, S. Chun, J. Choe, Y. Yoo, Cutmix: Regularization strategy to train
strong classifiers with localizable features, in: Proceedings of the IEEE/CVF international
conference on computer vision, 2019, pp. 6023–6032.
[27] H. Zhang, M. Cisse, Y. N. Dauphin, D. Lopez-Paz, mixup: Beyond empirical risk minimiza-
tion, arXiv preprint arXiv:1710.09412 (2017).
[28] K. He, X. Chen, S. Xie, Y. Li, P. Dollár, R. Girshick, Masked autoencoders are scalable vision
learners, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern
Recognition, 2022, pp. 16000–16009.
[29] J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, L. Fei-Fei, Imagenet: A large-scale hierarchical
image database, in: 2009 IEEE conference on computer vision and pattern recognition,
Ieee, 2009, pp. 248–255.
[30] A. K. Menon, S. Jayasumana, A. S. Rawat, H. Jain, A. Veit, S. Kumar, Long-tail learning via
logit adjustment, arXiv preprint arXiv:2007.07314 (2020).
[31] Y. Cui, M. Jia, T.-Y. Lin, Y. Song, S. Belongie, Class-balanced loss based on effective number
of samples, in: Proceedings of the IEEE/CVF conference on computer vision and pattern
recognition, 2019, pp. 9268–9277.
[32] Q. Diao, Y. Jiang, B. Wen, J. Sun, Z. Yuan, Metaformer: A unified meta framework for
fine-grained recognition, arXiv preprint arXiv:2203.02751 (2022).
[33] J. Wang, W. Zhang, Y. Zang, Y. Cao, J. Pang, T. Gong, K. Chen, Z. Liu, C. C. Loy, D. Lin,
Seesaw loss for long-tailed instance segmentation, in: Proceedings of the IEEE/CVF
Conference on Computer Vision and Pattern Recognition, 2021, pp. 9695–9704.
[34] T. Chen, S. Kornblith, M. Norouzi, G. Hinton, A simple framework for contrastive learning
of visual representations, in: International conference on machine learning, PMLR, 2020,
pp. 1597–1607.
[35] A. Van den Oord, Y. Li, O. Vinyals, Representation learning with contrastive predictive
coding, arXiv e-prints (2018) arXiv–1807.
[36] H. Bao, L. Dong, F. Wei, Beit: Bert pre-training of image transformers, arXiv preprint
arXiv:2106.08254 (2021).
[37] T.-Y. Lin, P. Goyal, R. Girshick, K. He, P. Dollár, Focal loss for dense object detection, in:
Proceedings of the IEEE international conference on computer vision, 2017, pp. 2980–2988.
[38] P.-Y. Chou, C.-H. Lin, W.-C. Kao, A novel plug-in module for fine-grained visual classifica-
tion, arXiv preprint arXiv:2202.03822 (2022).
[39] R. Borsodi, D. Papp, Incorporation of object detection models and location data into snake
species classification, in: Working Notes of CLEF 2021 - Conference and Labs of the
Evaluation Forum, 2021.
[40] G. Jocher, A. Stoken, A. Chaurasia, J. Borovec, NanoCode012, TaoXie, Y. Kwon, K. Michael,
L. Changyu, J. Fang, A. V, Laughing, tkianai, yxNONG, P. Skalski, A. Hogan, J. Nadar,
imyhxy, L. Mammana, AlexWang1900, C. Fati, D. Montes, J. Hajek, L. Diaconu, M. T. Minh,
Marc, albinxavi, fatih, oleg, wanghaoyang0106, ultralytics/yolov5: v6.0 - YOLOv5n ’Nano’
models, Roboflow integration, TensorFlow export, OpenCV DNN support, 2021. URL:
https://doi.org/10.5281/zenodo.5563715. doi:10.5281/zenodo.5563715.
[41] T. Chen, C. Guestrin, XGBoost: A scalable tree boosting system, in: Proceedings of the
22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining,
KDD ’16, ACM, New York, NY, USA, 2016, pp. 785–794. URL: http://doi.acm.org/10.1145/
2939672.2939785. doi:10.1145/2939672.2939785.
[42] A. Joly, H. Goëau, S. Kahl, L. Picek, T. Lorieul, E. Cole, B. Deneu, M. Servajean, A. Durso,
H. Glotin, R. Planqué, W.-P. Vellinga, A. Navine, H. Klinck, T. Denton, I. Eggel, P. Bonnet,
M. Šulc, M. Hruz, Overview of lifeclef 2022: an evaluation of machine-learning based
species identification and species distribution prediction, in: International Conference of
the Cross-Language Evaluation Forum for European Languages, Springer, 2022.
[43] M. Tan, Q. Le, Efficientnet: Rethinking model scaling for convolutional neural networks,
in: International Conference on Machine Learning, PMLR, 2019, pp. 6105–6114.