Evidence workspace

See what the video supports—and what still needs asking

Inspect speech and on-screen text alongside timecoded visual observations, attention, motion, and hand landmarks. Missing expert judgment is not inferred.

On-screen-text ledger · Missing analysis is not inferred

14

Speech + on-screen text

00

Direct visual observations

351

Visual annotations + derived cues

12

Real model artifacts

00

Interview probes

Ying Zhang / DONGTAI · 2:45 · fe6fba8e2329

Original source · Human review pending

Timecoded source evidence

14
Evidence 01
The direction can be adjusted

on screen text · human review pending

Evidence 02
Turn on, this is the light, this is the start button

on screen text · human review pending

Evidence 03
Speed adjustment

on screen text · human review pending

Evidence 04
First you need to grind the valve stem. Second, while grinding the valve stem, you need to grind the middle hole of the housing.

on screen text · human review pending

Evidence 05
Clamp the valve stem to be ground on the grinder

on screen text · human review pending

Evidence 06
Then select the ground inner bladder

on screen text · human review pending

Evidence 07
Corresponding to the thickness of the valve stem and fix it inside

on screen text · human review pending

Evidence 08
Then start the grinder and start moving forward or backward to grind

on screen text · human review pending

Evidence 09
This is a tapered rod

on screen text · human review pending

Evidence 10
When grinding, first fix the rod on the grinder and install the ring on the tapered rod

on screen text · human review pending

Evidence 11
First you need to measure its diameter and tap it in when it reaches the size you want

on screen text · human review pending

Evidence 12
Grind the middle hole of the shell and apply abrasive paste

on screen text · human review pending

Evidence 13
The valve stem must be flexible inside the housing

on screen text · human review pending

Evidence 14
After installation, measure the stroke and then go to the test bench to check the flow rate

on screen text · human review pending

02 · MULTIMODAL EVIDENCE

Visual evidence and machine observations

This separates literal video observations, authored attention, OpenCV-derived motion, and artifacts from models that actually ran. None of these alone establishes expert intent, rationale, quality, or safety.

Inspect overlays in Vision Lab
0Direct visual observations
351Visual annotations + derived cues
12Real model artifacts

Observations and model outputs support what is visible. What it means must be checked with a practitioner or separately cited existing knowledge.

01

Direct observations from the video

Literal descriptions of visible objects, orientation, possible contact, and movement only.

No direct visual observation has yet been admitted to this source bundle. The annotations and derived cues below remain visible, but they do not silently become expert knowledge.

02

Timecoded visual annotations and derived cues

Attention points, motion trails, hand landmarks, and related candidates. Play any card at its source moment.

03

Artifacts from visual models that actually ran

hand poseMediaPipe Hand Landmarker

Google · mediapipe-0.10.35:fbc2a30080c3

hand-clip-01.json
segmentationSAM 2.1 Hiera Tiny

Meta · facebookresearch/sam2@2b90b9f5ceec907a1c18123530e92e794ad901a4

keyframes.jsonclamp-lower-glove-mask.pngclamp-machine-chuck-mask.pngclamp-upper-glove-mask.pngclamp-valve-stem-mask.pnggrind-glove-mask.pnggrind-machine-chuck-mask.pnggrind-valve-stem-mask.png
segmentationSAM 2.1 Hiera Tiny video predictor

Meta · facebookresearch/sam2@2b90b9f5ceec907a1c18123530e92e794ad901a4

48000-58000-4fps.json50000-58000-4fps.json90000-100000-4fps.json

Continue with this source

One source, five connected views

Switching views preserves the same source hash across instruction, analysis, skill structure, evidence, and human review.