Papers › RADAM: Texture Recognition through Randomized Aggregated Encoding of Deep Activation Maps

RADAM: Texture Recognition through Randomized Aggregated Encoding of Deep Activation Maps

8 Mar 2023arXiv:2303.04554archive 2025-07-28

Leonardo Scabini, Kallil M. Zielinski, Lucas C. Ribas, Wesley N. Gonçalves, Bernard De Baets, Odemir M. Bruno

Texture analysis is a classical yet challenging task in computer vision for which deep neural networks are actively being applied. Most approaches are based on building feature aggregation modules around a pre-trained backbone and then fine-tuning the new architecture on specific texture recognition tasks. Here we propose a new method named \textbf{R}andom encoding of \textbf{A}ggregated \textbf{D}eep \textbf{A}ctivation \textbf{M}aps (RADAM) which extracts rich texture representations without ever changing the backbone. The technique consists of encoding the output at different depths of a pre-trained deep convolutional network using a Randomized Autoencoder (RAE). The RAE is trained locally to each image using a closed-form solution, and its decoder weights are used to compose a 1-dimensional texture representation that is fed into a linear SVM. This means that no fine-tuning or backpropagation is needed. We explore RADAM on several texture benchmarks and achieve state-of-the-art results with different computational budgets. Our results suggest that pre-trained backbones may not require additional fine-tuning for texture recognition if their learned representations are better encoded.

PaperPDFConference PDFCode

Code

scabini/RADAM officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

DecoderTexture Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification DTD RADAM (ConvNeXt-L) Accuracy 84.0 #2 of 11 Archive leaderboard report
Image Classification FMD (materials) RADAM (ConvNeXt-L) Accuracy (%) 95.2 #1 of 1 Archive leaderboard report
Image Classification KTH-TIPS2 RADAM (ConvNeXt-XL) Accuracy (%) 94.4 #1 of 1 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

RAERAdamSVM

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections