Eyes of the Future: Decoding the World Through Machine Vision

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. It is suggested to enhance the understanding of the shortcomings of machine vision in application processes during the transition from traditional methods to the latest advancements, as well as the challenges and innovations it faces.
2. In terms of future development, it is suggested that the author provide specific insights, such as how it will develop in the future and specific applications, etc.
3. To ensure the quality of the images in the research, some of them need to be processed with high quality rather than being taken directly as screenshots.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Clarify the methodology used for literature selection. Adding a systematic approach will strengthen the credibility of the review.

Recheck the reported performance values from the cited studies and ensure they are accurately reproduced. If values are unusually high, please provide context (controlled lab vs. real-world deployment). Otherwise, it risks overstating feasibility.

Add comparative discussion tables (e.g., performance trade-offs between neuromorphic vs. classical CNNs, edge vs. cloud processing). This will make the review more critical rather than descriptive.

Critically assess extraordinary claims (e.g., Mach 100 tracking) rather than restating them. The review should differentiate between proof-of-concept demonstrations and deployable systems.

Considr adding meta-level results (averaged findings, grouped outcomes by domain) rather than only listing single-study outcomes. This would better reflect the state of the art.

“Challenges and Limitations” section can be expanded to include practical deployment.

Organize future directions into a framework (short-term, mid-term, long-term trends), supported by specific methodological advances (e.g., benchmarks for quantum/neuromorphic systems).

Provide a clearer definition and maintain consistency when discussing MV vs. CV. Consider a diagram showing how MV fits within the broader CV/AI ecosystem.

Revise figures/tables to highlight comparative findings, trade-offs.

Comments on the Quality of English Language

Can be improved.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

the spike camera system (line 198 onward) is fascinating. that whole section about vform and SNN integration has strong potential. but the paragraph is long and dense. no breaks. also confusing that terms like “super vision” system are introduced (line 211) without explanation. coined by authors? a bit ambiguous. the description of the laser hitting the fan blades is vivid, but lacks scale or comparative metrics. how does this compare to existing systems? 

table formatting could really use some love. in Table 3 especially (line 337), categories feel overstuffed. for example, “Morphological Operations (Erosion, Dilation)” and “GANs” listed side-by-side, but they’re not comparable in scale or usage. also GANs listed twice in the same column (lines 334–335). maybe meant to be different uses but unclear. reader might assume it’s an error.
line 363 to 367 describes CNN feature extraction in a nice way but gets pretty textbook. nothing new here unless there’s an effort to tie back to earlier mentioned use-cases like plant stress or fruit grading. opportunity to show continuity missed.

the fiction references in autonomous decision section (lines 406–414) are unusual. yes, “Robbie” and Neuromancer are classics. but are they appropriate for a technical review paper? unless this journal has a history of science-literature analogies, those inclusions come off informal. maybe shift to footnotes if you really want to keep them.

later in line 859, “lack of diverse and sufficiently large datasets” is mentioned. okay, but in intro and abstract there was mention of “efficient data annotation techniques” (line 25). why not tie those together? a missed opportunity. need continuity between problems and solutions. 
section on quantum computing (line 789 onward) introduces SEQNN, QBC, AS, QSS, QNN, etc all back-to-back. a lot. hard to follow. maybe worth adding acronym definitions upfront or reduce some of the notation load. otherwise it’s cognitively fatiguing.

ethical concerns discussed briefly (line 893–902) but not really followed through. bias in training data, but what’s being done? if this is a review paper, needs to cite some concrete methods (e.g., adversarial debiasing, demographic parity metrics, etc.)

Author Response

Please see the attachment.

Author Response File: Author Response.pdf