On the Initialization of Swarm Intelligence Algorithms for Vector Quantization Codebook Design
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis is an interesting study and has a lot of interest for readers from many domains. The papers need a slight improvement before it can be accepted
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsAbstract needs to rewrite.
Authors should write contributions in bullet points in introduction.
Caption of figures should be written precisely. It is too long.
Few images are very small and low resolution. It will be difficult for reader to analyse.
Resolution must be at least 300 dpi of each figure.
What is the novelty of this work? Highlight in this paper.
How objective is attached should be technically explained.
Comments on the Quality of English LanguageFine.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe paper "Deep neural networks in simulating human visual illusions: Capabilities and limitations " is an important study of biological analogies of deep neural networks (DNN) in the field of vision through visual illusions. The authors discuss the limitations of current research, which are constrained by studying visual illusions in individual models on a single illusion. Further clarification is needed for a deeper understanding of visual illusions in DNN and expanding the types of illusions.
The results of this study are based on the analysis of five classic visual illusions in DNN, which demonstrated the capabilities and limitations of these networks in modeling human visual illusions. Significant differences in processing illusions based on attributes such as color, contrast, length, angle, and spatial positioning were noted.
The integration of DNN with human cognition is described. The authors emphasize that DNN exhibit human-like results in studying visual illusions, opening up new avenues for understanding the underlying neuro mechanisms. Integrating the study of visual illusions with DNN helps better understand the human visual system and advances the field of artificial intelligence.
The paper also highlights the limitations of DNN in accuracy and diversity of perception, pointing out fundamental differences between current models and the human brain in processing visual information. This underscores the need for further research.
The possibility of using DNN as a biological parallel analog for studying brain visual mechanisms is considered, although they should be viewed more as a model for study rather than a complete explanation of biological mechanisms.
These key points emphasize the significance of research in modeling visual illusions using deep neural networks, as well as indicating the prospects and limitations of this approach.
To improve the paper, the authors are recommended to:
- Propose new directions for further research based on the results obtained.
- Consider opportunities to expand the scope of research to other types of visual illusions.
The paper is recommended for publication as it represents an important study that significantly expands our understanding of the capabilities and limitations of deep neural networks in modeling human visual illusions. The results of the study, supported by RSA and CAM methods, are significant for the scientific community and can contribute to further development in the fields of artificial intelligence and neuroscience. The importance of the work also lies in identifying differences between DNN and human perception, aiding in better understanding the complexity of visual cognitive processes
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsComments are addressed