Muhammad
Zeeshan Javed

Problem

Golf coaches record full practice sessions, then scrub through hours of footage by hand to find each shot. Manual trimming does not scale past a few sessions.

Approach

A finite state machine watches the golfer through MediaPipe pose keypoints and flags candidate swings. YOLOv8 tracks the ball to confirm each candidate, and an audio impact detector validates the strike before a clip is cut.

Pipeline

Result

Fully automatic shot detection over long videos, delivered as a working pipeline.

Problem

COCO-trained detectors do not know what an auto-rickshaw is, and I could find no public Pakistan-specific mixed traffic dataset to train one on. Roads here mix rickshaws, bikes, carts and pedestrians in ways Western datasets never see.

Approach

Recorded 13 hours of dashcam footage from my own vehicle and hand-annotated 293 frames in VoTT with a 4-class taxonomy, split by source video to avoid leakage. Face detectors failed on hazy frames, so I built an anonymization pipeline around person detection with a full manual plate audit, released as its own open-source tool. Fine-tuned YOLO26n on a laptop CPU in under 3 hours and wired it into a proximity risk overlay with tracking and debounced labels.

Pipeline

Result

293 annotated frames, 1,506 boxes, 203 raw anonymized clips and the model, all public on Hugging Face. As a v0.1 daytime baseline on 54 validation images: mAP50 0.72 on original frames, 0.63 on the public anonymized ones.

Problem

Children lose interest in traditional online Quran classes, and parents have little visibility into their progress.

Approach

Built the academy web app through VisAItech: structured courses, free trial booking and live 1-to-1 classes, with levels and a mascot character to keep children engaged, plus progress tracking that parents can see.

Student journey

Result

Live at quranodyssey.com and taking free trial bookings.

Problem

Websites and apps for learning Pakistani Sign Language exist, but I could not find an AI-based learning solution. Producing signed video by hand needs interpreters and a studio for every sentence.

Approach

Recorded a word-level Pakistani Sign Language dataset in collaboration with two deaf-education institutions. Trained a two-stage pipeline: diffusion-based gloss to pose, then cGAN pose to video. Now migrating to a StableAnimator diffusion pipeline with DWPose conditioning and SeedVR2 enhancement.

Pipeline

Result

Paper under review at IEEE Access. Second-generation diffusion pipeline in progress.