TAPe+ML v3

TAPe+ML v3

Compact CV for classification, detection and segmentation

tapetape·Published on Nov 15, 2026·0 reputationReputation invested into this product to boost its ranking. Anyone can spend the reputation they've earned here — recent investments count for more and lift the product higher on the leaderboard. This is separate from community upvotes.
Visit siteOpen the product website in a new tab

▤ Screenshots

About TAPe+ML v3

TAPe+ML v3 is a compact computer vision architecture for image classification, object detection, and instance segmentation. It uses TAPe (Theory of Active Perception) representations rather than raw pixels or fixed image patches.

TAPe represents visual data as structured elements with known relationships. TAPe+ML v3 operates on these elements and their relationships through a set of small specialized submodels coordinated within one architecture. The system includes background separation, object localization, class prototype formation, and contour-based segmentation.

The paper reports 88.1% Top-1 accuracy on ImageNet-1k and 89.9% on ImageNet-ReaL with fewer than 100K parameters. On COCO object detection, TAPe+ML v3 reports 84.7 mAP50 and 65.3 mAP50-95 with fewer than 100K parameters. On COCO instance segmentation, it reports 80.7 Mask mAP50 and 58.4 Mask mAP50-95.

The paper also reports results on distribution-shift benchmarks: 78.6% on ObjectNet, 80.2% on ImageNet-v2, 90.3% on ImageNet-R, and 75.8% on ImageNet-Sketch.

On Imagenette, the same 3-layer CNN architecture trained under identical conditions reached 92% accuracy on TAPe data versus 47% on raw pixels. In a DINO/iBOT experiment, the standard architecture reached a loss of 0.406 on TAPe data with 9K examples, while the raw-pixel baseline did not converge with 120K examples.

TAPe was also evaluated for video scene detection. In the reported experiment, one hour of video was indexed in 10–11 seconds, the TAPe index was under 1 MB, and clustering took about one second.

The paper includes an industrial domain-shift pilot. With 30 images, YOLO26s achieved approximately 28% accuracy, TAPe+ML in head-only mode achieved approximately 45%, and TAPe+ML with backbone adaptation achieved approximately 65%. With 85 images, backbone adaptation reached approximately 85% accuracy. With 500 images and NMS, the reported range was 85.7–96%.

TAPe+ML also includes an auto-annotation workflow. After 10–15 images are labeled, the model can propose annotations for the remaining dataset; experts review and correct the proposed labels instead of annotating every object from scratch.

Read the arXiv preprint: https://arxiv.org/abs/2609.20869

Project page: https://ml.comexp.net

Contact: [email protected]

✦ Good to know

  • Train computer vision models on your own images or video for the object classes you need.
  • Deploy models in the cloud, on-premise, or directly on edge devices.
  • Start with ready-made COCO-pretrained models and a live detection demo before training your own.
  • Plans include a token-based ML workbench for testing recognition and creating custom models.

Taken from the product's own website and re-checked automatically.

DON'T TAKE OUR WORD FOR IT

Still not sure about TAPe+ML v3?

Let your favourite AI do the thinking. Pick one and see what it says.

Reviews

Loading reviews…

Comments

Loading comments…

Promoted

  1. Legal Chain
    Legal ChainPromoted
    Smart, fast, secure contracts from AI to blockchain.
    40 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote

More in AI

  1. Mnemosyne OS
    Mnemosyne OSmnemosyne-os.io
    Your AI memory, a control dashboard for humans and agents
    143 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
  2. MasterPlan AI
    MasterPlan AIapp.masterplanai.com.br
    AI-powered media planning and client management platform
    57 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
  3. Legal Chain
    Legal Chainlegalcha.in
    Smart, fast, secure contracts from AI to blockchain.
    40 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
  4. NotizPro
    NotizPronotiz.pro
    Capture voice notes and get instant AI summaries
    32 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
  5. Ray 3.2
    Ray 3.2ray32.net
    AI video generator creating cinematic clips from text or ima
    19 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
  6. husic
    husichusic.ai
    AI mastering that keeps your track sounding natural
    16 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
See all in AI →

You may also like

  1. Opsgenics
    Opsgenicsopsgenics.com
    DevOps and cloud development solutions for enterprise scalin
    16 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote
  2. TalkChief
    TalkChieftalkchief.io
    Unified communications platform with team messaging and AI
    12 PBSPBS (Points-Based Score) — a product's ranking score. It blends recent upvotes, saves, follows, ratings and real click-throughs with the reputation invested in the product, which decays like an upvote. Recent activity counts for more and scores fade as a product goes dormant, and no single signal can dominate. Higher PBS ranks higher on the leaderboard.
    Upvote