TechJam2026Track5

How the solution addresses the problem statement

The solution uses a lightweight convolutional neural network (ConvNeXt_Small) fused with radially averaged spectral cues, fine‑tuned on public datasets, together with an augmentation pipeline (color jitter, blur, crop, downsize and upsize, jpeg compression, flip and rotate) during training to simulate the post‑processing artifacts often encountered when images are shared on social media or messaging platforms.