EL MEHDI HICHAM

Computer Vision Engineer | Building TactiVision — 7 TensorRT engines, real-time football analytics | Ex-SportScore, Priori Sports | Morocco 🇲🇦

[EL MEHDI HICHAM @Mehdi_Hch_____]

⚽ Can AI see the next pass?

TactiVision Insight analyzes player positioning, passing lanes, blocked options and team shape directly from football video — turning movement into tactical intelligence.

🎥 Demo below.

#ComputerVision #SportsAI #FootballAnalytics #AI #SportsTech

[EL MEHDI HICHAM @Mehdi_Hch_____]

⚽ Can AI see the next pass?

TactiVision Insight analyzes player positioning, passing lanes, blocked options and team shape directly from football video — turning movement into tactical intelligence.

🎥 Demo below.

#ComputerVision #SportsAI #FootballAnalytics #AI #SportsTech

[EL MEHDI HICHAM @Mehdi_Hch_____]

⚽ Can AI see the next pass?

TactiVision Insight analyzes player positioning, passing lanes, blocked options and team shape directly from football video — turning movement into tactical intelligence.

🎥 Demo below.

#ComputerVision #SportsAI #FootballAnalytics #AI #SportsTech

[EL MEHDI HICHAM @Mehdi_Hch_____]

⚽ Can AI see the next pass?

TactiVision Insight analyzes player positioning, passing lanes, blocked options and team shape directly from football video — turning movement into tactical intelligence.

🎥 Demo below.

#ComputerVision #SportsAI #FootballAnalytics #AI #SportsTech

[EL MEHDI HICHAM @Mehdi_Hch_____]

⚽ Can AI see the next pass?

TactiVision Insight analyzes player positioning, passing lanes, blocked options and team shape directly from football video — turning movement into tactical intelligence.

🎥 Demo below.

#ComputerVision #SportsAI #FootballAnalytics #AI #SportsTech

[EL MEHDI HICHAM @Mehdi_Hch_____]

TactiVision ⚡️ TactiVision Insight — From Video → Data → Tactical Intelligence. Computer Vision is the foundation. Understanding the game is the next layer. ⚽️🧠

[EL MEHDI HICHAM @Mehdi_Hch_____]

53 seconds of 🇩🇪 Bayern Munich vs 🇫🇷 Paris Saint-Germain, analysed by TactiVision.
From tactical-camera footage to match KPIs, pass networks, team shape and player insights.

The sequence ended with Kvaratskhelia scoring from Doue's assist.

#TactiVision #ComputerVision #SportsAI

[EL MEHDI HICHAM @Mehdi_Hch_____]

53 seconds. One complete tactical report.
TactiVision analysed 🇹🇳 Tunisia vs 🇳🇱 Netherlands directly from video, generating match KPIs, pass networks, xT, team shape and player insights.

#TactiVision #FootballAnalytics #ComputerVision #SportsAI

[EL MEHDI HICHAM @Mehdi_Hch_____]

🏆#TactiVision — Champions League Stress Test
#PSG vs #Bayern
One tactical camera. Fast transitions, occlusions, tight marking and constant possession changes.
Different match. Same CV pipeline. No external tracking data
#ComputerVision #FootballAI #SportsAnalytics #TactiVision

[EL MEHDI HICHAM @Mehdi_Hch_____]

#Netherlands 🇳🇱 vs #Tunisia 🇹🇳
@FIFAWorldCup , rendered live.

Possession, territory, pressing, pass networks, heatmaps, pitch control — every number computed from the video alone.

🇳🇱 94% possession · 86% territory · 16 entries into danger areas
#TactiVision #ComputerVision

[EL MEHDI HICHAM @Mehdi_Hch_____]

Every player. Every frame. Automatically.

TactiVision turns broadcast footage into tracked identities — detection, re-ID, jersey OCR, real names. No manual tagging.

53 seconds of @Netherlands– @Tunisia

[EL MEHDI HICHAM @Mehdi_Hch_____]

The maps are not drawn They are what those trajectories produce once each player is projected onto a 105 x 68 m metric pitch:
→ every pass validated from possession changes and player movement
→ each one classified by type: progressive, through, switch, pressure

[EL MEHDI HICHAM @Mehdi_Hch_____]

One clip in. @FRMFOFFICIEL Morocco's passing map and activity heatmap out.

The video is the tracker: every player named from jersey OCR, speeds live.

The maps aren't drawn — they're what the trajectories produce once projected onto a metric pitch.

#TactiVision #ComputerVision

[EL MEHDI HICHAM @Mehdi_Hch_____]

🔗 Explore the full TactiVision project:
https://t.co/01oxPOiB2O
Explore the complete pipeline, technical architecture, results, and football analytics capabilities behind the demos.

[EL MEHDI HICHAM @Mehdi_Hch_____]

@17z_ai Gate, don’t interpolate. Ball behind bodies → explicit occluded state, no fabricated position. CV bridges 2-3 frames; in a box it’s 20+. Goal calls use 3-of-4 frame voting so dropout can’t break them

[EL MEHDI HICHAM @Mehdi_Hch_____]

🔗 Explore the full TactiVision project:
https://t.co/01oxPOiB2O
Explore the complete pipeline, technical architecture, results, and football analytics capabilities behind the demos.

[EL MEHDI HICHAM @Mehdi_Hch_____]

Harry Kane vs PSG. 17th minute.

TactiVision confirms the penalty:

→ Goal keypoints
→ Goal-line geometry
→ Ball tracking + temporal vote

Not "ball near goal."

Ball crossed reconstructed line inside valid goal-mouth.

Geometry, not detection.

#HarryKane #BayernMunich #PSG

[EL MEHDI HICHAM @Mehdi_Hch_____]

Morocco's goal against Haiti confirmed by AI in real time No Hawkeye. No goal-line technology. Just broadcast video + computer vision. #TactiVision reconstructs the goal geometry, tracks the ball polygon, and runs a temporal vote across consecutive frames
#Morocco #FootballAI

[EL MEHDI HICHAM @Mehdi_Hch_____]

Frame 237. Morocco vs Haiti.

Ball location: (38.2, 34.4m) in real-world pitch coordinates.
Homography error: 0.61m.
Running on a single RTX 4060.

Broadcast video → spatial intelligence. That's TactiVision.

#FootballAI #ComputerVision

[EL MEHDI HICHAM @Mehdi_Hch_____]

Every polished CV pipeline starts with a messy sketch on paper.
Here's how TactiVision detects whether the ball actually crossed the goal line — from notebook geometry to real-time inference 🧵
#TactiVision