Ruby, Red Pandas, and You
summary
A system identifies red pandas in long nest‑box videos by sampling one frame per second with FFMPEG, uploading with paperclip + jquery file upload (nginx upload_limit fixed), and classifying images using a TensorFlow Inception model with a 60% threshold plus three‑in‑a‑row smoothing. It converts 6–12 hour, 1+ GB videos into seconds‑indexed events, cuts human review time, and avoids ~$20,000 commercial costs; background: >50% captive cub mortality, <10,000 wild, 800–900 captive.
AI-written summary · about
chapters (25)
- 0:00Start
- 3:33BEE GUY
- 5:21RED PANDA TIME!
- 6:25WHAT IS A RED PANDA?
- 7:31RED PANDA DISTRIBUTION
- 7:51HABITAT AND ADAPTATIONS
- 8:46NUTRITIONAL ADAPTATIONS
- 9:00REPRODUCTIVE ADAPTATIONS
- 9:40CONSERVATION STATUS
- 10:07THREATS TO WILD POPS
- 10:48THREATS TO POPS UNDER HUMAN CARE
- 12:17BIRTH MONITORING
- 13:26MOTHER-CUB INTERACTIONS
- 13:46RESEARCH ON RED PANDAS
- 14:37USE OF TECHNOLOGY - NESTBOX CAMERAS
- 16:35STEP 1
- 17:32STEP 2
- 17:56MACHINE LEARNING
- 18:47CONVERSION
- 19:18STEP 4
- 20:29STEP 5
- 21:34RESULTS?
- 22:59WHAT IT ISN'T
- 23:53R4G GOALS
- 25:23COMMUNITY
related talks
28mChasing birds with code
37mUnstructured Data Processing with a Raspberry Pi AI Kit and Python
44mAdventures in Puppy Training with A.I. and a Raspberry Pi
21mLeveraging Python for Real-Time Image processing
29mRekognition Live Demo
31mYin Yang Ranch: Building a Distributed Computer Vision Pipeline using Python, O...
46mSaving elephants with Raspberry Pi’s & Machine Learning
31mExploring Real-time Computer Vision Using ActionCable