TDEC: Deep Embedded Image Clustering with Transformer and Distribution Information

Ruilin Zhang; Haiyang Zheng; Hongpeng Wang

cs.CVMar 23, 2026

TDEC: Deep Embedded Image Clustering with Transformer and Distribution Information

Authors: Ruilin Zhang, Haiyang Zheng, Hongpeng Wang

Organizations: Harbin Institute of Technology, Shenzhen, Shenzhen, China · Harbin Institute of Technology, Shenzhen, & Peng Cheng Laboratory, Shenzhen, China

Abstract

Image clustering is a crucial but challenging task in multimedia machine learning. Recently the combination of clustering with deep learning has achieved promising performance against conventional methods on high-dimensional image data. Unfortunately, existing deep clustering methods (DC) often ignore the importance of information fusion with a global perception field among different image regions on clustering images, especially complex ones. Additionally, the learned features are usually clustering-unfriendly in terms of dimensionality and are based only on simple distance information for the clustering. In this regard, we propose a deep embedded image clustering TDEC, which for the first time to our knowledge, jointly considers feature representation, dimensional preference, and robust assignment for image clustering. Specifically, we introduce the Transformer to form a novel module T-Encoder to learn discriminative features with global dependency while using the Dim-Reduction block to build a friendly low-dimensional space favoring clustering. Moreover, the distribution information of embedded features is considered in the clustering process to provide reliable supervised signals for joint training. Our method is robust and allows for more flexibility in data size, the number of clusters, and the context complexity. More importantly, the clustering performance of TDEC is much higher than recent competitors. Extensive experiments with state-of-the-art approaches on complex datasets show the superiority of TDEC.

Figures & tables

Figure 1. The framework of our proposed TDEC. It consists of T-Encoder, T-Decoder, Dim-Reduction block (DR), and Clustering Head (CH). Note that $L_{stru}$ and $L_{clu}$ represent the structure loss (the linear combination of reconstruction loss $L_{rec}$ and dimension reduction loss $L_{dim}$ ) and clustering loss.

Figure 2. An individual Transformer block.

Figure 3. The architecture of Dim-Reduction block. Here the features

Z^{i}_{w}

of size 110 learned by the T-Encoder for raw image $x_{i}$ is used as input to project a more cluster-friendly feature $Z^{i}_{v}$ of size 12.

Panel of Figure 4 — Figure 4. The convergence process of our model on MNIST

Dataset	# Samples	# size	# classes	# type
MNIST 1 1 1 http://yann.lecun.com/exdb/mnist/	70000	12828	10	Large-Scale
MNIST-test 1 1 footnotemark: 1	10000	12828	10	–
Fashion ( Xiao et al., 2017 )	70000	12828	10	Large-Scale
USPS ( Hull, 1994 )	9298	11616	10	–
Digits 2 2 2 https://archive.ics.uci.edu/ml/machine-learning-databases/optdigits/	1797	188	10	Small-scale
GTSRB 3 3 3 http://benchmark.ini.rub.de	8790	33232	5	Complex Contextual

Table 1. Dataset statistics.

Figure 5. Clustering samples on the YTF dataset.Each row contains the top 20 scoring images from one cluster, based on the distance from the cluster center.

Figure 6. Clustering samples of TDEC on GTSRB. Each row contains the top 20 scoring images from one cluster, based on the distance from the cluster center.

ST	CH	DR	MNIST	USPS	Digits	Letter	GTSRB
$\times$	$\times$	$\times$	ACC:0.84 NMI:0.81	ACC:0.74 NMI:0.75	ACC:0.52 NMI:0.52	ACC:0.53 NMI:0.51	ACC:0.66 NMI:0.55
✓	$\times$	$\times$	ACC:0.92 NMI:0.84	ACC:0.82 NMI:0.80	ACC:0.83 NMI:0.83	ACC:0.60 NMI:0.59	ACC:0.72 NMI:0.63
✓	✓	$\times$	ACC:0.93 NMI:0.86	ACC:0.90 NMI:0.83	ACC:0.87 NMI:0.80	ACC:0.65 NMI:0.67	ACC:0.74 NMI:0.68
✓	✓	✓	ACC: 0.97 NMI: 0.93	ACC: 0.97 NMI: 0.93	ACC: 0.95 NMI: 0.91	ACC: 0.68 NMI: 0.70	ACC: 0.82 NMI: 0.71

Table 4. Ablation study results of the proposed method and its degradation models.

Panel of Figure 7 — Figure 7. Output results of T-Encoder and T-Decoder on YTF dataset. (a): the raw images of three individuals in each of the 3 poses; (b): the corresponding reconstructed images.

Figure 8. Robustness of TDEC to different $k$ , which can be derived that 1) $k$ has good stability and 2) the neighborhood information is more favorable than the traditional distance information for clustering.

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Dataset	# Samples	# image size	# classes	# type
MNIST ( LeCun et al., 1998 )	70000	12828	10	Large-Scale
MNIST-test ( LeCun et al., 1998 )	10000	12828	10	Unbalanced
Fashion ( Xiao et al., 2017 )	70000	12828	10	Large-Scale
USPS ( Hull, 1994 )	9298	11616	10	Unbalanced
Digits ( Dua and Graff, 2017 )	1797	188	10	Small-scale
GTSRB ( Houben et al., 2013 )	8790	33232	5	Complex contextual

Appendix

Table 5. Dataset statistics.

Panel of Figure 9 — Figure 9. Datasets visualization.

Name	# source	# data source	# code source
JULE ( Yang et al., 2016 )	2016/CVPR	original or related papers	-
DEPICT ( Ghasedi Dizaji et al., 2017 )	2017/ICCV	original or related papers	-
DeepCluster ( Caron et al., 2018 )	2018/ECCV	original or related papers	-
CC ( Li et al., 2021 )	2021/AAAI	original or related papers	-
ASPC-DA ( Guo et al., 2020 )	2020/TKDE	our implementation	https://github.com/XifengGuo/ASPC-DA
VaGAN-SMM ( Yang et al., 2022 )	2022/TNNLS	original or related papers	-

Appendix

Table 6. Comparison algorithms statistics.

Clustering

Clustering groups data points by shared structure, developing methods to discover, validate, and interpret groups across diverse data and constraints.

See all 335 papers

Deep Image Clustering Based on Curriculum Learning and Density Information
Mar 31, 2026Haiyang Zheng, Ruilin Zhang, Hongpeng WangDensity-Based Clustering Clustering
Harbin Institute of Technology, Shenzhen, Shenzhen, China · Harbin Institute of Technology, Shenzhen & Peng Cheng Laboratory, Shenzhen, China
Leaner Transformers Can Easily Learn to Cluster
Oct 7, 2026Charlotte Park, Kenneth L. Clarkson, Lior Horesh +2Transformer Clustering
MIT · IBM Research
CAESAR: Clustering via Autonomous Embedding-Space Agglomerative Reorganization
Sep 26, 2026Ilan Bacry, Rémi Devaux, Antoine JardinCluster Number Estimation Clustering
Arlequin AI

Mar 31, 2026J/K move · Enter open · S save

Deep Image Clustering Based on Curriculum Learning and Density Information

Haiyang Zheng, Ruilin Zhang, Hongpeng Wang

Harbin Institute of Technology, Shenzhen, Shenzhen, China · Harbin Institute of Technology, Shenzhen & Peng Cheng Laboratory, Shenzhen, China

Density-Based Clustering Clustering Image Clustering Deep Clustering Curriculum Learning

Open paper AskPDF arXiv

Image clustering is one of the crucial techniques in multimedia analytics and knowledge discovery. Recently, the Deep clustering method (DC), characterized by its ability to perform feature learning and cluster assignment jointly, surpasses the performance of traditional ones on image data. However, existing methods rarely consider the role of model learning strategies in improving the robustness and performance of clustering complex image data. Furthermore, most approaches rely solely on point-to-point distances to cluster centers for partitioning the latent representations, resulting in error accumulation throughout the iterative process. In this paper, we propose a robust image clustering method (IDCL) which, to our knowledge for the first time, introduces a model training strategy using density information into image clustering. Specifically, we design a curriculum learning scheme grounded in the density information of input data, with a more reasonable learning pace. Moreover, we employ the density core rather than the individual cluster center to guide the cluster assignment. Finally, extensive comparisons with state-of-the-art clustering approaches on benchmark datasets demonstrate the superiority of the proposed method, including robustness, rapid convergence, and flexibility in terms of data scale, number of clusters, and image context.

Abstract

Clustering

Explore similar work

Deep Image Clustering Based on Curriculum Learning and Density Information

Leaner Transformers Can Easily Learn to Cluster

CAESAR: Clustering via Autonomous Embedding-Space Agglomerative Reorganization