Hybrid / FFT-Enhanced Transfer Baseline

[lukhsaankumar]

Issue Type

• Model: ML model bug, training issue, or architecture problem
• Data: Dataset issue, preprocessing bug, or data pipeline problem
• Web: Frontend bug or UI issue in the Next.js dashboard
• API: Backend API bug or FastAPI endpoint issue
• Research: Research question or experimental feature request
• Documentation: Documentation bug or improvement needed
• Bug: General bug fix needed
• Enhancement: New feature or improvement request

Description

• Combine a pretrained CNN or ViT backbone with the custom FFT feature branch. The backbone provides learned spatial/global features; the FFT branch adds frequency-domain artifacts. Concatenate both feature vectors and train a small classifier head end-to-end. (The FFT sub-network will be trained from scratch - no pretrained weights exist.)

Deliverable

• Notebook: `notebooks/hybrid_fft_transfer_baseline.ipynb
• Demonstrates fusion logic, training results, and early comparison to single-branch baselines

Additional Context

Pretrained Model Options:

• CNN Backbone: EfficientNet-B0 or ResNet-50 (same as Issue CNN Transfer Baseline #22 )
• ViT Backbone: ViT-Base Patch16 224 or DeiT-Small Patch16 224 (same as Issue ViT Transfer Baseline #23 )

[lukhsaankumar]

Put this ticket on pause for now and work on Issue #27