Chocolate-dip chromis · w32_fish4

Chromis dimidiata · Pomacentridae · 91% · high reconstructed model: good

Reviewer: two-tone body, forked tail

Where it comes from

Original video (longest visit, 34:21-34:27), the selected fish boxed

Tracking: real frame + crop + 3D render at the estimated pose, species overlay

3D and 4D model

loading 3D gaussians…

Interactive canonical 3D gaussians (drag to orbit, scroll to zoom). This is the rest pose shared by every frame; the 4D part is the per-frame bending / tail motion shown on the right.

4D: canonical model replaying the learned body bending and tail beat

3D turntable of the canonical model

Poster

Species

Chocolate-dip chromis Chromis dimidiata Pomacentridae

90.7%

Axilspot hogfish Bodianus axillaris Labridae

6.8%

Lyretail hogfish Bodianus anthioides Labridae

0.7%

Bicolour parrotfish Cetoscarus bicolor Scaridae

0.7%

Singlebar devil Plectroglyphidodon leucozonus Pomacentridae

0.6%

Family: Pomacentridae 92%, Labridae 8%, Scaridae 1% · agreement across crops 92% · 64 crops · bioclip-2.5 (hf-hub:imageomics/bioclip-2.5-vith14, open_clip, zero-shot)

Data and reconstruction

window	32:00 – 40:00
visits	11: 32:08-32:10, 33:12-33:14, 34:22-34:26, 34:57-34:59, 35:06-35:08, 35:40-35:42, 36:08-36:09, 36:19-36:21, 36:38-36:40, 39:53-39:53, 39:56-39:58
frames (10 fps) detected / kept	218 / 194
valid poses (10 fps)	195
frames with background plate	193
median size	118.0 px
30 fps frames / valid poses	395 / 303
method	4D GS, fins + bending + tail (30 fps)
model shown	v1 (gs_final)
iterations	4000
gaussians	15481
training time	20 min
trainer validation PSNR full / masked	26.9 / 19.9 dB
trainer validation IoU vs SAM mask	0.840
near-fish band PSNR (fins)	27.9 dB
production status	done
visual note (curation)	Chromis dimidiata likely

Diagnostics

Held-out frames: real crop vs 4D render (composited over the plate)

Estimated pose: fitted ellipsoid + axes on crops

Frame cleaning: identity / occlusion over time

Frames down-weighted / dropped by the robust trainer

Downloads

point_cloud.ply (canonical 3D gaussians) · poses_refined.json · model.pt (incl. per-frame 4D params) · species.json

← all fish