Transaction-Science/open-standards
Watch
1
0
Fork
You've already forked open-standards
 0
Code Issues Pull requests Projects Releases 15 Packages Wiki Activity Actions

wai-rs: jpegai_icci — the Haar wavelet domain of the eICCI post-filter #13

Merged
dcharlot merged 1 commit from feat/jpegai-dwt into main 2026-07-15 03:26:53 -04:00
Conversation 0 Commits 1 Files changed 2 +175
dcharlot commented 2026-07-15 03:26:52 -04:00
Owner
Copy link

First piece of eICCI, the last unported JPEG-AI decode stage. Its CNN operates in a wavelet domain rather than on pixels; this lands that transform.

A Haar wavelet on 2x2 blocks with 1/2 scaling, two levels: [1,H,W] → [4,H/2,W/2] → [16,H/4,W/4]. Stacking the three planes gives the 48 channels the network first conv consumes.

LL= a+b+c+d ; HL=-a-b+c+d ; LH=-a+b-c+d ; HH= a-b-c+d   (a..d pre-scaled by 1/2)

Perfect reconstruction: LL-HL-LH+HH = 4a, so the inverse recovers each sample exactly. Channel layout matches the reference cat order.

Unit tests: hand-checked 2x2 block, idwt∘dwt = id, two-level round-trip + 16-band shape, and the 48-channel stack.

First piece of **eICCI**, the last unported JPEG-AI decode stage. Its CNN operates in a wavelet domain rather than on pixels; this lands that transform. A Haar wavelet on 2x2 blocks with 1/2 scaling, two levels: `[1,H,W]` → `[4,H/2,W/2]` → `[16,H/4,W/4]`. Stacking the three planes gives the **48 channels** the network first conv consumes. ``` LL= a+b+c+d ; HL=-a-b+c+d ; LH=-a+b-c+d ; HH= a-b-c+d (a..d pre-scaled by 1/2) ``` Perfect reconstruction: `LL-HL-LH+HH = 4a`, so the inverse recovers each sample exactly. Channel layout matches the reference cat order. Unit tests: hand-checked 2x2 block, `idwt∘dwt = id`, two-level round-trip + 16-band shape, and the 48-channel stack.
First piece of eICCI, the last unported JPEG-AI decode stage. eICCI's CNN operates in
a wavelet domain rather than on pixels; this lands that transform.

A Haar wavelet on 2x2 blocks with 1/2 scaling, applied two levels: a plane [1,H,W]
becomes [4,H/2,W/2] then [16,H/4,W/4]. Stacking the three planes' subbands gives the
48 channels the network's first convolution consumes (cat(Y16,U16,V16)).

  LL =  a+b+c+d    (all pre-scaled by 1/2)
  HL = -a-b+c+d
  LH = -a+b-c+d
  HH =  a-b-c+d

Perfect reconstruction: LL-HL-LH+HH = 4a, so the inverse recovers each sample exactly.
Channel layout matches the reference cat order — one level emits [LL,HL,LH,HH]; two
levels emit x16[i*4..i*4+4] = the level-2 decomposition of level-1 band i.

Unit tests: hand-checked single 2x2 block, idwt.dwt = id, two-level round-trip with the
16-band shape, and the 48-channel plane stack.

Co-Authored-By: Claude Opus 4.8 <[email protected]>
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
Transaction-Science/open-standards!13
No description provided.