Alignment is crucial for text-to-image (T2I) models to ensure that the generated images faithfully capture user intent while maintaining safety and fairness. Direct Preference Optimization (DPO) has emerged as a key alignment technique for large language models (LLMs), and its influence is now extending to T2I systems. This paper introduces DPO-Kernels for T2I models, a novel extension of DPO that enhances alignment across three key dimensions: (i) Hybrid Loss, which integrates embedding-based objectives with the traditional probability-based loss to improve optimization; (ii) Kernelized Representations, leveraging Radial Basis Function (RBF), Polynomial, and Wavelet kernels to enable richer feature transformations, ensuring better separation between safe and unsafe inputs; and (iii) Divergence Selection, expanding beyond DPO’s default Kullback–Leibler (KL) regularizer by incorporating alternative divergence measures such as Wasserstein and Rényi divergences to enhance stability and robustness in alignment training. We introduce DETONATE, the first large-scale benchmark of its kind, comprising approximately 100K curated image pairs, categorized as chosen and rejected. This benchmark encapsulates three critical axes of social bias and discrimination: Race, Gender, and Disability. The prompts are sourced from the hate speech datasets, while the images are generated using state-of-the-art T2I models, including Stable Diffusion 3.5 Large (SD-3.5), Stable Diffusion XL (SD-XL), and Midjourney. Furthermore, to evaluate alignment beyond surface metrics, we introduce the Alignment Quality Index (AQI) for T2I systems: a novel geometric measure that quantifies latent space separability of safe/unsafe image activations, revealing hidden model vulnerabilities. While alignment techniques often risk overfitting, we empirically demonstrate that DPO-Kernels preserve strong generalization bounds using the theory of Heavy-Tailed Self-Regularization (HT-SR).

Open-Vocabulary semantic segmentation aims to segment images into distinct semantic regions for both seen and unseen categories at the pixel level. Current methods utilize text embeddings from pre-trained vision-language models like CLIP but struggle with the inherent domain gap between image and text embeddings, even after extensive alignment during training. Additionally, relying solely on deep text-aligned features limits shallow-level feature guidance, which is crucial for detecting small objects and fine details, ultimately reducing segmentation accuracy. To address these limitations, we propose a dual prompting framework, DPSeg, for this task. Our approach combines dualprompt cost volume generation, a cost volume-guided decoder, and a semantic-guided prompt refinement strategy that leverages our dual prompting scheme to mitigate alignment issues in visual prompt generation. By incorporating visual embeddings from a visual prompt encoder, our approach reduces the domain gap between text and image embeddings while providing multi-level guidance through shallow features. Extensive experiments demonstrate that our method significantly outperforms existing state-of-the-art approaches on multiple public datasets. The code and models are available: https://github.com/ziyuz-vision/DPSeg.

Abstract Dense connections in convolutional neural networks (CNNs), which connect each layer to every other layer, can compensate for mid/high‐frequency information loss and further enhance high‐frequency signals. However, dense CNNs suffer from high memory usage due to the accumulation of concatenating feature‐maps stored in memory. To overcome this problem, a two‐step approach is proposed that learns the representative concatenating feature‐maps. Specifically, a convolutional layer with many more filters is used before concatenating layers to learn richer feature‐maps. Therefore, the irrelevant and redundant feature‐maps are discarded in the concatenating layers. The proposed method results in 24% and 6% less memory usage and test time, respectively, in comparison to single‐image super‐resolution (SISR) with the basic dense block. It also improves the peak signal‐to‐noise ratio by 0.24 dB. Moreover, the proposed method, while producing competitive results, decreases the number of filters in concatenating layers by at least a factor of 2 and reduces the memory consumption and test time by 40% and 12%, respectively. These results suggest that the proposed approach is a more practical method for SISR.

Facial dynamics can be considered as unique signatures for discrimination between people. These have started to become important topic since many devices have the possibility of unlocking using face recognition or verification. In this work, we evaluate the efficacy of the transition frames of video in emotion as compared to the peak emotion frames for identification. For experiments with transition frames we extract features from each frame of the video from a fine-tuned VGG-Face Convolutional Neural Network (CNN) and geometric features from facial landmark points. To model the temporal context of the transition frames we train a Long-Short Term Memory (LSTM) on the geometric and the CNN features. Furthermore, we employ two fusion strategies: first, an early fusion, in which the geometric and the CNN features are stacked and fed to the LSTM. Second, a late fusion, in which the prediction of the LSTMs, trained independently on the two features, are stacked and used with a Support Vector Machine (SVM). Experimental results show that the late fusion strategy gives the best results and the transition frames give better identification results as compared to the peak emotion frames.