Drag-based editing within pretrained diffusion model provides a precise and flexible way to manipulate foreground objects. Traditional methods optimize the input feature obtained from DDIM inversion directly, adjusting them iteratively to guide handle points towards target locations. However, these approaches often suffer from limited accuracy due to the low representation ability of the feature in motion supervision, as well as inefficiencies caused by the large search space required for point tracking. To address these limitations, we present DragLoRA, a novel framework that integrates LoRA (Low-Rank Adaptation) adapters into the drag-based editing pipeline. To enhance the training of LoRA adapters, we introduce an additional denoising score distillation loss which regularizes the online model by aligning its output with that of the original model. Additionally, we improve the consistency of motion supervision by adapting the input features using the updated LoRA, giving a more stable and accurate input feature for subsequent operations. Building on this, we design an adaptive optimization scheme that dynamically toggles between two modes, prioritizing efficiency without compromising precision. Extensive experiments demonstrate that DragLoRA significantly enhances the control precision and computational efficiency for drag-based image editing. The Codes of DragLoRA are available at: https://github.com/Sylvie-X/DragLoRA.

Testing autonomous driving systems (ADS) presents significant challenges in trajectory planning, route control, and collision avoidance, particularly at complex junctions. Among these, irregular junctions are especially valuable for exposing ADS weaknesses—yet they remain difficult to generate systematically. Since manually configured irregular junctions may not accurately reflect real-world conditions, a more practical approach is to identify existing irregular junctions. This paper presents a method for extracting irregular junctions from global OpenStreetMap (OSM) data and generating safety-critical scenarios based on them. By analyzing junction topologies and quantifying them with a difficulty metric, our approach uncovers challenging scenarios that effectively expose ADS defects. Experimental evaluations with various autopilots show that the identified junctions and generated scenarios trigger unique ADS defects more effectively than simulator-provided ones.

ComOpT is an open-source research tool for coverage-driven testing of autonomous driving systems, focusing on planning and control. Starting with (i) a meta-model characterizing discrete conditions to be considered and (ii) constraints specifying the impossibility of certain combinations, ComOpT first generates constraint-feasible abstract scenarios while maximally increasing the coverage of k-way combinatorial testing. Each abstract scenario can be viewed as a conceptual equivalence class, which is then instantiated into multiple concrete scenarios by (1) randomly picking one local map that fulfills the specified geographical condition, and (2) assigning all actors accordingly with parameters within the range. Finally, ComOpT evaluates each concrete scenario against a set of KPIs and performs local scenario variation via spawning a new agent that might lead to a collision at designated points. We use ComOpT to test the Apollo 6 autonomous driving software stack. ComOpT can generate highly diversified scenarios with limited test budgets while uncovering problematic situations such as inabilities to make simple right turns, uncomfortable accelerations, and dangerous driving patterns. ComOpT participated in the 2021 IEEE AI Autonomous Vehicle Testing Challenge and won first place among more than 110 contending teams.

This study proposes a novel super‐resolution regularisation model based on adaptive sparse representation and self‐learning frameworks. The fidelity term in the model ensures that the reconstructed image is consistent with the observation image. The adaptive sparsity regularisation term constrains the reconstructed image with an adaptive sparse representation, which successfully harmonises the sparse representation and the collaborative representation adaptively via producing suitable coefficients. To construct a more effective dictionary, the high‐frequency features from the underlying image patches are extracted, and the dictionary learning and sparse representation are integrated. To this end, the alternating minimisation algorithm is used to divide this model into three subproblems, and the alternating direction method of multipliers and iterative back‐projection method are used to solve the subproblems. To illustrate the effectiveness of the proposed method, additional experiments are conducted on some generic images. Compared with some state‐of‐the‐art algorithms, the experimental results demonstrate that the proposed method achieves better results in terms of both visual quality and noise immunity.

In this paper, we apply outer synchronization to identify unknown parameters existing in the node dynamics between two interacted networks. According to the coupling and interacted matrices, we design the corresponding adaptive control schemes and updating laws to achieve the outer synchronization. By the Lyapunov functional theory, we derive two theorems of the appearance of outer synchronization. Finally we provide numerical simulations to demonstrate the validity of our theoretical results and show the process of outer synchronization and the identification of unknown parameters.

The progressive fine granularity scalable (PFGS) coding is a promising technique for streaming video applications. The macroblock-based PFGS coding proposes several INTER modes for the enhancement macroblock coding and provides a simple method to determine a coding mode for each macroblock. Because it does not consider the drifting error possibly occurs in decoder, the mode decision method is not effective. In this paper, we present a mode decision scheme to make a trade-off between the low drifting error and high coding efficiency, by adding a middle bit-rate loop in encoder to estimate the possible drifting error in decoder and using an R-D model to estimate the improvement of coding efficiency by selecting high quality reference. Then a nearly optimal coding mode for each enhancement macroblock can be easily determined. The experimental results show that the proposed scheme can effectively reduce the drifting error at low bit rates, while maintaining improved coding efficiency at high bit rates

Wavelet-based scalable video coding is a promising technique for streaming video applications; especially the motion compensated temporal filtering (MCTF). It usually decomposes the input video frames into several subbands in temporal domain. A quantization error in each subband contributes to the overall distortion in the ultimate reconstructed video with different weights. These weighting factors play an important role in allocating bit budget to different subbands and influence the coding performance greatly. In theory, these weighting factors can be derived from the wavelet filter, but they are not match the experimental data very well. This paper first proposes a probability model to estimate the influence of different motion connections in MCTF. And the zero-coefficient percentage of each subband is presented to model the influence of accumulated rounding error. These two proposed models can make a good estimation on the subband weighting factors. The experimental results show that they have the consistent trend with actual data.