=Paper=
{{Paper
|id=Vol-2595/endoCV2020_Hu_Guo_et_al
|storemode=property
|title=Endoscopic Artefact Detection in MMDetection
|pdfUrl=https://ceur-ws.org/Vol-2595/endoCV2020_paper_id_s29.pdf
|volume=Vol-2595
|authors=Hongyu Hu,Yuanfan Guo
|dblpUrl=https://dblp.org/rec/conf/isbi/HuG20
}}
==Endoscopic Artefact Detection in MMDetection==
https://ceur-ws.org/Vol-2595/endoCV2020_paper_id_s29.pdf
ENDOSCOPIC ARTEFACT DETECTION IN MMDETECTION
Hongyu Hu1 , Yuanfan Guo2
1
Hongyu Hu , Shanghai Jiaotong University, mathewcrespo@sjtu.edu.cn
2
Yuanfan Guo , Shanghai Jiaotong University, gyfastas@sjtu.edu.cn
Table 1. Baseline performance on validation data set
AP AP IoU =.50 AP IoU =.75 AP small AP medium AP large
1. METHODS 0.260 0.514 0.228 0.060 0.127 0.323
1.1. Architecture
We use Cascade-RCNN [1], which is a multi-stage object de- Table 2. Performance on validation data set with multi-scale
tection architecture as our base model and adopt ResNeXt [2] detection
AP AP IoU =.50 AP IoU =.75 AP small AP medium AP large
as backbone with Feature Pyramid Networks (FPN) [3] for
0.277 0.539 0.250 0.068 0.152 0.335
feature extraction.
1.2. Implement details Table 3. Results on 100% test data set with different parame-
ters
• Mmdetection toolbox Mmdetection [4] is toolbox for threshold 0.030 0.030 0.050 0.050 0.100 0.100 0.200 0.200
object detection with many state-of-the-art and pre- max 100 20 100 20 100 20 100 20
dscore 0.184 0.194 0.189 0.195 0.116 0.215 0.2115 0.2202
trained models, which is very practical in this task.
• Data augmentation Each image has 50 percent chance
to be flipped horizontally. Table 4. Final result on 100% test data set
Score d dscore dstd gmAP gdev
• Soft-nms We use soft-nms [5] rather than nms to avoid 0.2202±0.0562 0.2202 0.0562 0.1671 0.0879
objects being directly ignored by mistake. We carry
out a series of experiments on soft-nms threshold and
maximum number of bounding boxes to better avoid by 0.008, as is shown in Table 2. Notably, the boost of AP
over-detected objects. mainly comes from performance on medium and large ob-
jects. We infer that medium and large objects are also zoomed
• Multi-scale detection Test images and training images out and the model has better global cognition over the image.
are of different scales. When training, images are re-
sized randomly from (512, 512) to (1024, 1024). We
2.2. Trade-off on bounding box’s number
are able to have a closer look on small objects.
In given training data set and test data set, each image mainly
2. RESULTS has about few to tens of bounding boxes [6][7][8]. When
inference, threshold in soft-nms and maximum number of
We use 4/5 of the data set for training and the rest for evalua- bounding boxes in each image decide the number of bound-
tion. ing boxes. In Table 3, we list experiment results on this pair
of parameters and decide threshold and maximum number set
2.1. Object detection of different sizes as 0.2 and 20.
As baseline result is shown in Table 1, AP small is much
smaller than AP medium and AP large . Accurate detection for 2.3. Final result
small object is the bottleneck of this task. After introducing We mainly use multi-scale detection and proper parameter
multi-scale detection, performance on small objects improves settings in soft-nms to solve the problems mentioned above.
Copyright c 2020 for this paper by its authors. Use permitted under Final result on 100 % test set is shown in Table 4. This result
Creative Commons License Attribution 4.0 International (CC BY 4.0). ranks 8th in final leader board.
3. REFERENCES
[1] Zhaowei Cai and Nuno Vasconcelos. Cascade r-cnn:
High quality object detection and instance segmentation.
arXiv preprint arXiv:1906.09756, 2019.
[2] Saining Xie, Ross Girshick, Piotr Dollr, Zhuowen Tu, and
Kaiming He. Aggregated residual transformations for
deep neural networks. arXiv preprint arXiv:1611.05431,
2016.
[3] Tsung-Yi Lin, Piotr Dollár, Ross B. Girshick, Kaiming
He, Bharath Hariharan, and Serge J. Belongie. Fea-
ture pyramid networks for object detection. CoRR,
abs/1612.03144, 2016.
[4] Kai Chen, Jiaqi Wang, Jiangmiao Pang, Yuhang Cao,
Yu Xiong, Xiaoxiao Li, Shuyang Sun, Wansen Feng,
Ziwei Liu, Jiarui Xu, Zheng Zhang, Dazhi Cheng,
Chenchen Zhu, Tianheng Cheng, Qijie Zhao, Buyu Li,
Xin Lu, Rui Zhu, Yue Wu, Jifeng Dai, Jingdong Wang,
Jianping Shi, Wanli Ouyang, Chen Change Loy, and
Dahua Lin. MMDetection: Open mmlab detection tool-
box and benchmark. arXiv preprint arXiv:1906.07155,
2019.
[5] Navaneeth Bodla, Bharat Singh, Rama Chellappa, and
Larry S. Davis. Soft-nms – improving object detection
with one line of code. 2017.
[6] Sharib Ali, Felix Zhou, Barbara Braden, Adam Bai-
ley, Suhui Yang, Guanju Cheng, Pengyi Zhang, Xiao-
qiong Li, Maxime Kayser, Roger D. Soberanis-Mukul,
Shadi Albarqouni, Xiaokang Wang, Chunqing Wang,
Seiryo Watanabe, Ilkay Oksuz, Qingtian Ning, Shufan
Yang, Mohammad Azam Khan, Xiaohong W. Gao, Ste-
fano Realdon, Maxim Loshchenov, Julia A. Schnabel,
James E. East, Geroges Wagnieres, Victor B. Loschenov,
Enrico Grisan, Christian Daul, Walter Blondel, and Jens
Rittscher. An objective comparison of detection and seg-
mentation algorithms for artefacts in clinical endoscopy.
Scientific Reports, 10, 2020.
[7] Sharib Ali, Felix Zhou, Christian Daul, Barbara Braden,
Adam Bailey, Stefano Realdon, James East, Georges
Wagnieres, Victor Loschenov, Enrico Grisan, et al. En-
doscopy artifact detection (EAD 2019) challenge dataset.
arXiv preprint arXiv:1905.03209, 2019.
[8] Sharib Ali, Felix Zhou, Adam Bailey, Barbara Braden,
James East, Xin Lu, and Jens Rittscher. A deep learning
framework for quality assessment and restoration in video
endoscopy. arXiv preprint arXiv:1904.07073, 2019.