pith. sign in
Pith Number

pith:MLJD76SL

pith:2026:MLJD76SLC5IM3FGFNJSMMOZBPV
not attested not anchored not stored refs resolved

FaSST: Fast Sparsifying Secondary Transform

Antonio Ortega, Darukeesan Pakiyarajah, Debargha Mukherjee, Eduardo Pavez, Samuel Fern\'andez-Mendui\~na

FaSST factorizes data-driven secondary transforms into mode-adaptive Givens rotations to match LFNST performance at far lower complexity.

arxiv:2605.15086 v1 · 2026-05-14 · eess.IV · eess.SP

Add to your LaTeX paper
\usepackage{pith}
\pithnumber{MLJD76SLC5IM3FGFNJSMMOZBPV}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more

Record completeness

1 Bitcoin timestamp
2 Internet Archive
3 Author claim open · sign in to claim
4 Citations open
5 Replications open
Portable graph bundle live · download bundle · merged state
The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Mode-adaptive FaSST matches the RD performance of LFNST while reducing the number of computations by 83.67%. Moreover, by avoiding fixed-coefficient truncation, FaSST achieves up to 1.80% BD-rate savings relative to LFNST while operating at 66.24% lower complexity.

C2weakest assumption

That the alternating-minimization procedure combined with approximate Givens factorization produces transforms whose rate-distortion performance remains close to the original data-driven SOTs across the full range of intra prediction modes and content types used in AV2 testing.

C3one line summary

FaSST approximates sparse orthonormal transforms with mode-adaptive Givens rotation sequences to produce low-complexity secondary transforms for AV2 intra residuals that match LFNST rate-distortion performance at 83.67% lower complexity and deliver up to 1.80% BD-rate savings.

References

31 extracted · 31 resolved · 2 Pith anchors

[1] FaSST: Fast Sparsifying Secondary Transform 2026 · arXiv:2605.15086
[2] coefficient dropping
[3] Moreover, KLTs and their approx- imations do not account for quantization effects
[4] given the coefficients{ ˆyi}me i=1, we optimize the transform: min SJ meX i=1 ∥ˆxi −S J ˆyi∥2 2 s.t.S J = JY j=1 G(mj, nj, θj).(6) It can be shown that the above problem is equivalent to max SJ tr(ΓSJ
[5] Specifically, images from the CLIC dataset [20] are compressed using A V2, and residual blocks of sizes 8×8,16×16, and32×32are collected

Formal links

1 machine-checked theorem link

Cited by

1 paper in Pith

Receipt and verification
First computed 2026-05-17T21:40:25.953441Z
Last reissued 2026-05-17T21:57:19.265601Z
Builder pith-number-builder-2026-05-17-v1
Signature unsigned_v0
Schema pith-number/v1.0

Canonical hash

62d23ffa4b1750cd94c56a64c63b217d7b1515c2ca3930b1b4b07f35d8e88265

Aliases

arxiv: 2605.15086 · arxiv_version: 2605.15086v1 · pith_short_12: MLJD76SLC5IM · pith_short_16: MLJD76SLC5IM3FGF · pith_short_8: MLJD76SL
Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/MLJD76SLC5IM3FGFNJSMMOZBPV \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 62d23ffa4b1750cd94c56a64c63b217d7b1515c2ca3930b1b4b07f35d8e88265
Canonical record JSON
{
  "metadata": {
    "abstract_canon_sha256": "e707a7a8d3972a515746bf548f4366c9700f094b5ddc9d65d9e5d489498bdab7",
    "cross_cats_sorted": [
      "eess.SP"
    ],
    "license": "http://creativecommons.org/licenses/by-nc-nd/4.0/",
    "primary_cat": "eess.IV",
    "submitted_at": "2026-05-14T17:08:16Z",
    "title_canon_sha256": "64425464d993df606cc7fc6ed995be1bdfc2981b5bd8b3935f2df2fce82a1915"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.15086",
    "kind": "arxiv",
    "version": 1
  }
}