What is it and how quickly you can guess?
Żaneta Demarczyk, Katarzyna Spyra, Adrian Trojanowski
Faculty of Applied Mathematics, Silesian University of Technology
Kaszubska 23, 44-100 Gliwice, Poland
zaneta.demarczyk93@gmail.com, katarzyna.spyra@interia.pl, adtrojanowski@interia.pl
Abstract—This article presents research on human between robots and autonomous systems. All these ideas are
interactions by using a methodology of gradually revealed images helpful in the research on human behavior. In [9] was
for recognition. The idea we measure here it to compare results discussed how to use a composition of neural networks and
of interactions while guessing on the image. In the results we heuristic methods to detect some features of fruits from
show and discuss differences between sex of the participant and
category of the quiz.
images, while in [10] was proposed a method for automatic
selection of bacteria. On the other hand there are many
Keywords—image composition, interactions, human behavior, research on object oriented programming where cognitive
sociology of decisions aspects are modeled to increase code efficiency. In [11]
authors proposed some complexity metrics based on cognitive
I. INTRODUCTION models, while in [12] were presented research results on
Interactions are driven by many factors. During decision reactions to vocalization of dogs and their emotional aspects.
processes our brain is focusing on some aspects of the reality Results of using human behavioral models are very important
which can be easily associated with the things we have in our for autonomous systems, where groups of unmanned robots
memory or which surround us. This interactions are driven by are set to perform complex tasks, but communication between
some factors which we can associate and use for conclusions. them is based on human behaviors. In [13] was discussed how
Very often we must decide under pressure or under limited to model a self-organizing strategies for autonomous group of
time. For these we can find some differences between man and robots in changing environments, while in [14] these were
woman, since not only a brain but also a sociology of decision compared to performance of interactions between working
is important. There are many articles presenting results from agents. The aim of this project is to show interactions between
decision processes, where humans were asked to describe human and computer. For this reason we have developed a
reactions from various inputs like sounds, images, unexpected program which presents images to users and measures their
situations, etc. In [1] was presented how humans react to the choices basing on the category. The program takes the form of
sound of aircraft. Authors measured reactions and described a game and selects one of the available images from given
them in relation to the user. In [2] was presented how humans field and shows a part of randomly chosen pixels. The number
react to rewards and punishments in various situations, where of pixels which are discovered increases with passing time,
as an exemplary social model was realized theory of Gray’s until all pixels are shown and whole picture is presented. In
personality. In [3] was presented how humans react to this time user is asked to guess what in his opinion is
uncontrolled results of situations they participate in, the presented in the revealed image. In our program we have three
authors were especially interested on relation of interactions to available categories: buildings, famous people and animals. Of
superstition. Very often in the research on human interactions course, user knows the categories, but he doesn’t see the
are used images. From an image we are able to evaluate many images in advance. In every field there are 5 pictures, which
emotions and also knowledge about the content. In [4] were are selected randomly. In our opinion, such games have a very
discussed reactions to images, eg. by facial or behavioral good effect on people. They examine perceptiveness and
features. In [5] authors discussed both reactions to images and knowledge from various fields (eg from geography, history)
also motivations that were diving people to interact in each therefore we have decided to present some research results in
way. In [6] were presented differences between man and this field of human interactions to images.
woman reactions to children facial images, while in [7]
differences were discussed on example of animals. An II. DETAILS OF THE PROGRAM
interesting aspects of psychological tendencies in our brains This project is written in Wolfram Mathematica 10 for
during choices were discussed in [8]. Image processing and research purposes. Now, we talk a bit about the code. In the
interactions between machines based on human behavior are program, we used the fact that every image can be presented
widely discussed in recent times. New articles present as a pixel’s matrix. The algorithm randomly selects and show
interesting ideas for selecting objects from images or to used from 5% to 50% of pixels of each row with the step 5%. At
models of human interactions to proceed communications the last stage the whole picture is exposed. Sample
visualization of the process is presented in Fig. 1.
Copyright held by the author(s).
17
Fig. 1 A sample sequence of the images during quiz shown starting from 5%, while the user is asked to guess what is presented in the image.
18
Fig. 2 Part of the code of the program in Mathematica 10 student edition.
Tab. 1 Results obtained from 20 players.
Observation number Sex Category Picture Time [s] Observation number Sex Category Picture Time [s]
buildings Sphinx 24,2173 buildings Colosseum 19,2601
1 Male 11 Female
people Karol Wojtyła 0 people Enrique Iglesias 21,3165
animals Elephant 12,575 animals Squirrel 26,093
buildings Tower of Pisa 5,35274 buildings Sphinx 18,4756
2 Male 12 Female
people Robert Lewandowski 12,6457 people Robert Lewandowski 16,7513
animals Squirrel 34,4054 animals Flamingo 15,46
buildings Tower of Pisa 19,3325 buildings Eiffel Tower 0
3 Male 13 Female
people Enrique Iglesias not guessed people Rihanna 15,0165
animals Gorilla 28,9175 animals Squirrel 23,51922
buildings Statue of Liberty 0 buildings Eiffel Tower 3,9064
4 Male 14 Male
people Robert Lewandowski 21,3615 people Marilyn Monroe 0
animals Flamingo 16,7 animals Elephant 7,71992
buildings Sphinx not guessed buildings Tower of Pisa 5,9608
5 Female 15 Male
people Enrique Iglesias 26,6507 people Marilyn Monroe 4,71504
animals Squirrel 33,4138 animals Elephant 6,647
buildings Colosseum 29,9062 buildings Colosseum not guessed
6 Female 16 Female
people Marilyn Monroe 4,14567 people Rihanna 9,1617
animals Horse 25,9206 animals Horse 19,8639
buildings Eiffel Tower 4,48744 buildings Eiffel Tower 0
7 Female 17 Male
people Marilyn Monroe 7,78586 people Enrique Iglesias 31,05
animals Gorilla 23,0975 animals Flamingo 17,4206
buildings Colosseum 23,14 buildings Eiffel Tower 5,61302
8 Female 18 Female
people Rihanna 8,40952 people Enrique Iglesias 20,2053
animals Elephant 9,15681 animals Elephant 5,21907
buildings Tower of Pisa 0 buildings Tower of Pisa 0
9 Male 19 Female
people Rihanna not guessed people Karol Wojtyła 12,564
animals Horse 23,6915 animals Squirrel 22,5378
buildings Eiffel Tower 0 buildings Statue of Liberty 13,3166
10 Male 20 Male
people Karol Wojtyła 8,47303 people Robert Lewandowski 15.2067
animals Horse 24,60212 animals Squirrel 20,7439
19
Tab. 2 Results obtained from 20 players.
Whole
Number of observations The shortest time [s] Average time [s] The longest time [s]
Everything 60 0 14,18525341 35
Buildings 20 0 12,148435 35
Sphinx 3 18,4756 25,89763333 35
Tower of Pisa 5 0 6,129208 19,3325
Eiffel Tower 6 0 2,334476667 5,61302
Statue of Liberty 2 0 6,6583 13,3166
Colosseum 4 19,2601 26,826575 35
People 20 0 15,27643789 35
Robert Lewandowski 4 12,6457 16,9195 21,3615
Karol Wojtyła 3 0 7,012343333 12,564
Rihanna 4 8,40952 16,89693 35
Enrique Iglesias 5 20,2053 26,8445 35
Marilyn Monroe 4 0 4,1616425 7,78586
Animals 20 5,21907 19,885232 34,4054
Elephant 5 5,21907 8,26356 12,575
Horse 4 19,8639 23,51953 25,9206
Flamingo 3 15,46 16,52686667 17,4206
Squirrel 6 20,7439 26,78552 33,4138
Gorilla 2 23,0975 26,0075 28,9175
Tab. 3 Results for male participants.
Male
Number of observations The shortest time [s] Average time [s] The longest time [s]
Everything 30 0 14,26739828 35
Buildings 10 0 7,208634 24,2173
Sphinx 1 24,2173 24,2173 24,2173
Tower of Pisa 4 0 7,66151 19,3325
Eiffel Tower 3 0 1,302133333 3,9064
Statue of Liberty 2 0 6,6583 13,3166
Colosseum 0
People 10 0 16,47169667 35
Robert Lewandowski 3 12,6457 17,0036 21,3615
Karol Wojtyła 2 0 4,236515 8,47303
Rihanna 1 35 35 35
Enrique Iglesias 2 31,05 33,025 35
Marilyn Monroe 2 0 2,35752 4,71504
Animals 10 6,647 19,342294 34,4054
Elephant 3 6,647 8,98064 12,575
Horse 2 23,6915 24,14681 24,60212
Flamingo 2 16,7 17,0603 17,4206
Squirrel 2 20,7439 27,57465 34,4054
Gorilla 1 28,9175 28,9175 28,9175
20
Tab. 4 Results for female participants.
Female
Number of observations The shortest time [s] Average time [s] The longest time [s]
Everything 30 0 17,239037 35
Buildings 10 0 17,088236 35
Sphinx 2 18,4756 26,7378 35
Tower of Pisa 1 0 0 0
Eiffel Tower 3 0 3,36682 5,61302
Statue of Liberty 0
Colosseum 4 19,2601 26,826575 35
People 10 4,14567 14,200705 26,6507
Robert Lewandowski 1 16,7513 16,7513 16,7513
Karol Wojtyła 1 12,564 12,564 12,564
Rihanna 3 8,40952 10,86257333 15,0165
Enrique Iglesias 3 20,2053 22,72416667 26,6507
Marilyn Monroe 2 4,14567 5,965765 7,78586
Animals 10 5,21907 20,42817 33,4138
Elephant 2 5,21907 7,18794 9,15681
Horse 2 19,8639 22,89225 25,9206
Flamingo 1 15,46 15,46 15,46
Squirrel 4 22,5378 26,390955 33,4138
Gorilla 1 23,0975 23,0975 23,0975
Buildings
40
35
30
25
Time
20
Whole
15
Male
10
Female
5
0
The shortest time [s] Average time [s] The longest time [s]
Whole 0 12,148435 35
Male 0 7,208634 24,2173
Female 0 17,088236 35
Fig. 3 Comparison of results in category buildings.
21
Animals
40
35
30
25
Time
20
Whole
15
Male
10 Female
5
0
The shortest time [s] Average time [s] The longest time [s]
Whole 5,21907 19,885232 34,4054
Male 6,647 19,342294 34,4054
Female 5,21907 20,42817 33,4138
Fig. 4 Comparison of results in category animals.
People
40
35
30
25
Time
20
Whole
15
Male
10
Female
5
0
The shortest time [s] Average time [s] The longest time [s]
Whole 0 15,27643789 35
Male 0 16,47169667 35
Female 4,14567 14,200705 26,6507
Fig. 5 Comparison of results in category people.
22
Female
40
35
30
Time 25
20
15 Buildings
10
5 People
0 Animals
The shortest Average time The longest
time [s] [s] time [s]
Buildings 0 17,088236 35
People 4,14567 14,200705 26,6507
Animals 5,21907 20,42817 33,4138
Fig. 6 Comparison of results due to sex of players.
Male
40
35
30
25
Time
20
15 Buildings
10 People
5 Animals
0
The shortest The longest time
Average time [s]
time [s] [s]
Buildings 0 7,208634 24,2173
People 0 16,47169667 35
Animals 6,647 19,342294 34,4054
Fig. 7 Comparison of results due to sex of players.
The player who correctly guessed with the shortest time wins,
A. What does the user do?
but we know that sometimes it is unfair because of the
At the beginning, the user has to select the category of pictures difficulty of the several images. The algorithm is running as
by simply typing one of commands. long as the matrix is filled with the proper amount of the
pixels without duplicates. In each step algorithm works from
beginning- it means that it’s not picking the missing quantity
Then the image is being exposed with the passing time. When of pixels to the matrix from the previous step. The pictures in
the user already knows what the picture shows he/she should Fig. 1 show the next stages of the program’s work on a
push “PAUSE”. The part of our code is shown in Fig. 2. randomly selected image.
23
III. RESULTS References
We invited 20 people to play our game. Everyone tried his
[1] Kryter, Karl D. "Scaling human reactions to the sound from aircraft."
chances in each category and we received the results presented The Journal of the Acoustical Society of America 31.11 (1959): 1415-
in Tab. 1 – Tab. 4 and depicted in Fig. 3 – Fig. 7: 1429.
[2] Wilson, Glenn D., Paul T. Barrett, and Jeffrey A. Gray. "Human
For example. The sixth user was woman and the shortest time reactions to reward and punishment: A questionnaire examination of
Gray's personality theory." British Journal of Psychology 80.4 (1989):
she obtained was in the category – people. She guessed that in 509-515.
the picture was Marlin Monroe. [3] Matute, Helena. "Human reactions to uncontrollable outcomes: Further
evidence for superstitions rather than helplessness." The Quarterly
It can be observed that during our tests the most often Journal of Experimental Psychology Section B 48.2b (1995): 142-157.
displayed building was Eiffel Tower, in category people – [4] Lang, Peter J., et al. "Looking at pictures: Affective, facial, visceral, and
Enrique Iglesias and in category animals it was squirrel. The behavioral reactions." Psychophysiology 30.3 (1993): 261-273.
shortest average time needed to guess was for buildings. [5] Beritelli, Francesco, Capizzi, Giacomo, Lo Sciuto, Grazia, Napoli,
Christian and Scaglione, Francesco. “Automatic heart activity diagnosis
based on gram polynomials and probabilistic neural
Pictures that presented Marlin Monroe and elephant have got networks”. Biomedical Engineering Letters, 8.1 (2018): 77-85.
the shortest time in their categories. [6] Bradley, Margaret M., et al. "Emotion and motivation I: defensive and
appetitive reactions in picture processing." Emotion 1.3 (2001): 276.
We can see that differences between man and woman on the [7] Platek, Steven M., et al. "Reactions to children's faces: Resemblance
charts visible in Fig. 4 – Fig. 5. In the category buildings, men affects males more than females." Evolution and Human Behavior 23.3
(2002): 159-166.
turned out to be better. In the category people better were
[8] Bovet, Dalila, and Jacques Vauclair. "Picture recognition in animals and
women. The smallest differences between men’s and women’s humans." Behavioural brain research 109.2 (2000): 143-165.
time was in the category animals. [9] Levy, Sidney J. "Dreams, fairy tales, animals, and cars." Psychology &
Marketing 2.2 (1985): 67-81.
IV. CONCLUSION [10] Woźniak, Marcin, Dawid Połap, Christian Napoli, and Emiliano
Generally men have shorter time than women. Among men, Tramontana. "Application of bio-inspired methods in distributed gaming
systems." Information Technology And Control 46, no. 1 (2017): 150-
the shortest average time is for pictures from category 164.
buildings. While by the decisions all the users were most [11] Woźniak, Marcin, and Dawid Połap. "Adaptive neuro-heuristic hybrid
convenient with images of nature and calm colors, and where model for fruit peel defects detection." Neural Networks 98 (2018): 16-
the colors were strict and very light these images were not 33.
very convenient to users. Sometimes the images were [12] Starczewski, Janusz T., Sebastian Pabiasz, Natalia Vladymyrska,
correctly identified in first 5 second due to some explicit Antonino Marvuglia, Christian Napoli, and Marcin Woźniak. "Self
organizing maps for 3D face understanding." In International
details visible in presented objects. On the other hand these Conference on Artificial Intelligence and Soft Computing, (2016): pp.
were not much visible for people. In general users were 210-217. Springer.
correctly recognizing people in images when some facial [13] Woźniak, Marcin, et al. "Automated fluorescence microscopy image
details appeared and were not able to recognize in first analysis of Pseudomonas aeruginosa bacteria in alive and dead stadium."
Engineering Applications of Artificial Intelligence 67 (2018): 100-110.
seconds when only a shape was visible.
[14] Misra, Sanjay, et al. "A Suite of Object Oriented Cognitive Complexity
Metrics." IEEE Access (2018).
In our opinion the research gave us important clues how the [15] Maskeliunas, Rytis, Vidas Raudonis, and Robertas Damasevicius.
people react to various objects. These conclusions will be very "Recognition of Emotional Vocalizations of Canine." Acta Acustica
useful in our future work, where we can use them for united with Acustica 104.2 (2018): 304-314.
implementing systems oriented communication aspects, where [16] Esmaeili, Ahmad, et al. "A socially-based distributed self-organizing
a recognition of the input will be determined by some initial algorithm for holonic multi-agent systems: Case study in a task
environment." Cognitive Systems Research 43 (2017): 21-44.
information about the input objects.
[17] Esmaeili, Ahmad, et al. "The impact of diversity on performance of
holonic multi-agent systems." Engineering Applications of Artificial
Intelligence 55 (2016): 186-201.
24