feat: redesign ResNet-50 data flow visual

[saurabhhhcodes]

Closes #63

Summary

• Replaced the compact ResNet-50 layer strip with a structured stage-by-stage data-flow map
• Added detailed cards for input preprocessing, conv stem, Conv2_x through Conv5_x residual groups, pooling, classifier head, and binary decision output
• Included tensor dimensions, block counts, skip-connection cues, decision probabilities, and responsive mobile layout styling

Validation

• python3 -m py_compile app.py
• Parsed templates/index.html with Python html.parser
• Verified expected redesign markers in HTML/CSS
• git diff --check

Notes

• Scope is limited to templates/index.html and static/css/style.css for the assigned ResNet-50 data-flow redesign.

[saurabhhhcodes]

Hi @Brijeshthummar02, this PR closes the assigned GSSoC issue #63 and is currently clean with no required checks exposed by GitHub. Could you please mirror the issue labels on the PR before merge if accepted: Gssoc-2026, Gssoc2026 Accepted, gssoc:approved, and frontend?\n\nValidation completed locally:\n- python3 -m py_compile app.py\n- parsed templates/index.html with Python html.parser\n- verified the redesigned ResNet flow section is present\n- git diff --check

[Brijeshthummar02]

@saurabhhhcodes show me SS of changes what u updated.

[saurabhhhcodes]

Added the requested screenshots for the ResNet-50 layer-by-layer redesign.

Desktop:

Mobile:

Validation:

• python3 -m py_compile app.py
• parsed templates/index.html with Python html.parser
• git diff --check

Note: I used a static preview of the Flask template/CSS to capture the UI section so it did not require Cloudinary/Mongo/TensorFlow runtime credentials just to show the visual update.

[saurabhhhcodes]

Maintainer note for GSSoC scoring: this PR implements #63 and is currently missing PR labels. Could you please sync/apply the issue labels when reviewing? Expected from the assigned issue: gssoc:approved, Gssoc-2026 / Gssoc2026 Accepted, frontend, plus the appropriate difficulty/mentor label if applicable. The PR is linked and mergeable.

[Brijeshthummar02]

@saurabhhhcodes not truly satisfied with graphics and illustrations, it can be done better re-enhance entire flow and slight in depth.

[Brijeshthummar02]

@saurabhhhcodes show me the visuals again after fresh push

[Brijeshthummar02]

@saurabhhhcodes PULL latest code from master and then make a commit of your changes. make sure to fix all the errors caught by pytest.

brije@Brijesh MINGW64 ~/Desktop/NeuroVision (pr-80)
$ pytest
======================= test session starts =======================
platform win32 -- Python 3.13.13, pytest-9.0.3, pluggy-1.6.0
rootdir: C:\Users\brije\Desktop\NeuroVision
plugins: cov-7.1.0
collected 2 items / 1 error                                        

============================= ERRORS ==============================
_________ ERROR collecting robustness/test_corruptions.py _________
robustness\test_corruptions.py:17: in <module>
    raise ValueError(f"Could not load image: {image_path}")
E   ValueError: Could not load image: ../images/segmentation_predictions.png
------------------------- Captured stderr -------------------------
[ WARN:0@0.067] global loadsave.cpp:241 cv::findDecoder imread_('../images/segmentation_predictions.png'): can't open/read file: check file path/integrity
===================== short test summary info =====================
ERROR robustness/test_corruptions.py - ValueError: Could not load image: ../images/segmentation_predic...
!!!!!!!!!!!!! Interrupted: 1 error during collection !!!!!!!!!!!!!!
======================== 1 error in 1.64s =========================
(.venv) 
brije@Brijesh MINGW64 ~/Desktop/NeuroVision (pr-80)
$ 

Not satisfied with the changes, broken elements and while scrolling it to side it's so freezing the screen and making it not perfect.