Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Joachim WeickertSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works62 from public data
- 360
- Normalized Cut Loss for Weakly-Supervised CNN Segmentation351
This work proposes a new principled loss function evaluating network output with criteria standard in "shallow" segmentation, e.g. normalized cut which evaluates only seeds where labels are known while normalized cut softly evaluates consistency of all pixels.
- On Regularized Losses for Weakly-supervised CNN Segmentation339
This approach simplifies weakly-supervised training by avoiding extra MRF/CRF inference steps or layers explicitly generating full masks, while improving both the quality and efficiency of training.
- A Fully Progressive Approach to Single-Image Super-Resolution285
To obtain more photorealistic results, a generative adversarial network (GAN) is designed, named ProGanSR, that follows the same progressive multi-scale design principle and constitutes a principled multi- scale approach that increases the reconstruction quality for all upsampling factors simultaneously.
- 218
- PhaseNet for Video Frame Interpolation200
This work proposes a new approach, PhaseNet, that is designed to robustly handle challenging scenarios while also coping with larger motion, and shows that this is superior to the hand-crafted heuristics previously used in phase-based methods and compares favorably to recent deep learning based approaches for video frame interpolation on challenging datasets.
- High‐Resolution Neural Face Swapping for Visual Effects153
This is the first method capable of rendering photo‐realistic and temporally coherent results at megapixel resolution and incorporates a refinement strategy into the face landmark stabilization algorithm to achieve temporal stability, which is crucial for working with high‐resolution videos.
- 110
- Blind image super-resolution with spatially variant degradations98
This work proposes a solution that relies on a degradation aware SR network to synthesize the HR image given a low resolution image and the corresponding blur kernel and presents an optimization procedure that is able to recover both the degradation kernel and the high resolution image by minimizing the error predicted by the kernel discriminator.
- Cyclic Schemes for PDE-Based Image Analysis68
This work investigates a class of efficient numerical algorithms for many partial differential equations (PDEs) in image analysis that are applicable to parabolic or elliptic PDEs that have bounded coefficients and lead to space discretisations with symmetric matrices and shows that one should use parameter cycles that result from factorisations of box filters.
- Deep Generative Video Compression63
This work proposes an end-to-end, deep generative modeling approach to compress temporal sequences with a focus on video that builds upon variational autoencoder models for sequential data and combines them with recent work on neural image compression.
- VideoSnapping60
This work first derives a robust approximation of alignment quality between pairs of clips, computed as a weighted histogram of feature matches, and finds optimal temporal mappings (constituting frame correspondences) using a graph-based approach that allows for very efficient evaluation with artist constraints.
- 37
- Neural frame interpolation for rendered content34
This paper proposes a method specifically targeted to achieve high-quality frame interpolation for rendered content and is able to significantly push the state-of-the-art in frame interpolated motion estimates, more accurate occlusion handling and to correctly reconstruct non-linear motion between keyframes.
- Neural Video Compression with Spatio-Temporal Cross-Covariance Transformers33
This work proposes a new 3D-based transformer module: Spatio-Temporal Cross-Covariance Transformer (ST-XCT), which is integrated into various key coding components of NVC, such as feature extraction, frame reconstruction, and entropy modeling, demonstrating its generalizability.
- Controllable Inversion of Black-Box Face Recognition Models via Diffusion31
This method is the first black-box face recognition model inversion method that offers intuitive control over the generation process and leverages the stochastic nature of the denoising diffusion process to produce high-quality, identity-preserving face images with various backgrounds, lighting, poses, and expressions.
- An Omnistereoscopic Video Pipeline for Capture and Display of Real-World VR30
A first complete guide and analysis for reimplementing a system for capturing and displaying real-world VR video content, based on the concept of omnistereoscopic panoramas is provided.
- Anisotropic Range Image Integration27
This paper extends the state-of-the-art approach of Zach et al.(2007) in several ways, replacing the isotropic space-variant smoothing behaviour by an anisotropic (direction-dependent) one and using the more accurate closest signed distances instead of directional signed distances when converting range images into 3D signed distance fields.
- Deep Probabilistic Video Compression24
This model uses advances in variational autoencoders (VAEs) for sequential data and combines it with recent work on neural image compression to propose a variational inference approach to deep probabilistic video compression.
- 23
- Deep Adaptive Sampling and Reconstruction Using Analytic Distributions21
This work proposes an adaptive sampling and reconstruction method for offline Monte Carlo rendering that produces sampling maps constrained by a user-defined budget that minimize the expected future denoising error and supports joint end-to-end training of the sampling and Denoising networks.
- 20
- 20
- 18
- Introducing Maximal Anisotropy into Second Order Coupling Models17
This work proposes a second order regulariser that allows to penalise both jumps and kinks in a direction-dependent way, and systematically introduces anisotropic concepts from first order approaches.
- 16
- Versatile Vision Foundation Model for Image and Video Colorization14
This work shows how a latent diffusion model, pre-trained on text-to-image synthesis, can be finetuned for image colorization and provide a flexible solution for a wide variety of scenarios: high quality direct colorization with diverse results, user guided colorization through colors hints, text prompts or reference image and finally video colorization.
- Bridging the Gap between Gaussian Diffusion Models and Universal Quantization for Image Compression13
This work proposes a novel quantization-based forward diffusion process with theoretical foundations that tackles all three critical gaps in previous approaches following this paradigm, and achieves the best rate-distortion-realism performance, outperforming previous related works.
- High-Fidelity Novel View Synthesis via Splatting-Guided Diffusion13
This paper introduces SplatDiff, a pixel-splatting-guided video diffusion model designed to synthesize high-fidelity novel views from a single image, and proposes an aligned synthesis strategy for precise control of target viewpoints and geometry-consistent view synthesis.
- 13
- 13
- Reenact Anything: Semantic Video Motion Transfer Using Motion-Textual Inversion12
The method, called motion-textual inversion, leverages the observation that image-to-video models extract appearance mainly from the (latent) image input, while the text/image embedding injected via cross-attention predominantly controls motion, and represents motion using text/image embedding tokens.
- Kernel-Based Frame Interpolation for Spatio-Temporally Adaptive Rendering12
An adaptive strategy that allows predicting full or partial keyframes that should be rendered with color samples solely based on the auxiliary features of a shot and allows rendering significantly fewer color pixels as compared to a fixed-step scheme when wanting to maintain a certain quality.
- Learning Dynamic 3D Geometry and Texture for Video Face Swapping11
By leveraging the learned 3D geometry and texture, this method achieves face swapping with higher quality than when using off‐the‐shelf monocular 3D face reconstruction, and overall lower FID score than state‐of-the‐art 2D methods.
- 11
- Physically inspired depth-from-defocus10
A forward operator that preserves essential physical properties such as a maximumminimum principle for the intensity values is designed, which allows us to approximate the thin-lens camera model in a better way than previous approaches.
- Introducing More Physics into Variational Depth–from–Defocus8
This work presents a novel forward operator that approximates the thin–lens camera model from physics better than previous ones used for this task, since it preserves the maximum–minimum principle w.r.t. the unknown image intensities.
- Controllable Tracking-Based Video Frame Interpolation7
This work introduces a tracking-based video frame interpolation method that utilizes sparse point tracks, first estimated and interpolated with existing point tracking methods and then optionally refined by the user, and proposes a mechanism for controlling the levels of hallucination in interpolated frames through inference-time model weight adaptation.
- Multiview Depth Parameterisation with Second Order Regularisation7
This paper investigates several parameterisations of the unknown depth and reveals that an inverse depth parameterisation is generally preferable, and points out its benefits w.r.t. the data and the smoothness term.
- A Generative Model for Digital Camera Noise Synthesis6
This work proposes an effective generative model which utilizes clean features as guidance followed by noise injections into the network and shows that the proposed approach outperforms existing methods for synthesizing camera noise.
- 5
- 5
- A Dense Pipeline for 3D Reconstruction from Image Sequences5
A novel pipeline for 3D reconstruction from image sequences that solely relies on dense methods, which yields accurate camera poses as well as 3D reconstructions and is an interesting alternative to sparse approaches.
- 4
- 4
- Microdosing: Knowledge Distillation for GAN Based Compression4
This paper demonstrates how to leverage knowledge distillation to obtain equally capable image decoders at a fraction of the original number of parameters and shows how to transfer the obtained benefits into the setting of video compression.
- 3
- Cross Anisotropic Cost Volume Filtering for Segmentation3
This paper presents a substantially improved cost computation scheme which incorporates texture descriptors, as well as an automatic feature selection strategy which leads to a high-dimensional feature space, from which the label costs are extracted using a support vector machine.
- StableDPT: Temporal Stable Monocular Video Depth Estimation2
This work proposes a novel approach that adapts any state-of-the-art image-based image-based (depth) estimation model for video processing by integrating a new temporal module - trainable on a single GPU in a few days.
- 2
- Leveraging Diffusion Models for Stylization Using Multiple Style Images2
This work designs a method that leverages both image prompt adapters and statistical alignment of the features during the denoising process, and achieves state-of-the-art results for stylization.
- Revitalizing Legacy Video Content: Deinterlacing with Bidirectional Information Propagation2
A deep-learning-based method for deinterlacing animated and live-action content using a Flow-guided Refinement Block which performs feature refinement including alignment, fusion, and rectification and can process multiple fields simultaneously, reducing per-frame processing time, and potentially enabling real-time processing.
- 1
- 1
- 1
- 1
- 1
- Variational surface reconstruction1
A novel pipeline for 3D reconstruction from image sequences that solely relies on dense ideas and enables us to derive novel approaches to surface reconstruction from oriented points that are fairly simple and offer state-of-the-art performance.
- 2D Gaussian Splatting for Bézier Spline Line Art Vectorization–
This paper shows that it is possible to go beyond heuristics to extract a meaningful set of strokes from a raster image, and achieves state-of-the-art results for image reconstruction, while being fast and producing strokes of high quality.
- –
- Training a Deep Remastering Model–
This work presents a deep learning solution to bring the NTSC version to the new scan quality levels, which would be otherwise impossible with existing tools.
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.