Best Paper Awards

On the Versatile Uses of Partial Distance Correlation in Deep Learning

Best Paper Awards – Honorable Mention

A Level Set Theory for Neural Implicit Evolution under Explicit Flows

Pose-NDF: Modelling Human Pose Manifolds with Neural Distance Fields

Everingham Prize

Outstanding long-term service to the computer vision community

The UCF101 and HMD51 dataset teams

Koenderink Prize

A naturalistic open source movie for optical flow evaluation

Indoor Segmentation and Support Inference from RGBD Images

Best Demo

Using a Smartphone for Augmented Reality in a Classroom

Best Paper Awards

RAFT: Recurrent All-Pairs Field Transforms for Optical Flow

Best Paper Awards – Honorable Mention

Towards Streaming Perception

NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis

Everingham Prize

Multiple Datasets

COLMAP SFM and MVS software library

Koenderink Prize

Improving the Fisher Kernel for Large-Scale Image Classification

Brief: Binary robust independent elementary features

Best Demo

Inter-Homines

Best Paper Awards

Implicit 3D Orientation Learning for 6D Object Detection from RGB Images

Best Paper Awards – Honorable Mention

Group Normalization

GANimation: Anatomically-aware Facial Animation from a Single Image

Everingham Prize

For a series of datasets and workshops since 2003 that have driven progress in large scale Video Retrieval.

For providing a well documented software library for Structure from Motion that has been used effortlessly by so many.

Koenderink Prize

Hamming Embedding and Weak Geometric Consistency for Large Scale Image Search

Semi-supervised On-Line Boosting for Robust Tracking

Best Paper Awards

Real-Time 3D Reconstruction and 6-DoF Tracking with an Event Camera

Best Paper Awards – Honorable Mention

The Fast Bilateral Solver

Everingham Prize

For a series of datasets and challenges since 2010 that have had such impact on the computer vision field. ImageNet built on the Caltech101/256 datasets, increasing the number of images by orders of magnitude and enabling the development of new algorithms.

For extensive, generous, service to the community: As long-term head of the IEEE PAMI Technical Committee he introduced many reforms, including to the awards process and the relationship to the IEEE. And he has been the driving force in creating and running the Computer Vision Foundation (CVF).

Koenderink Prize

Surf: Speeded up robust features (ECCV 2006)

Machine learning for high-speed corner detection (ECCV 2006)

Best Student Paper Award

Focal Flow: Measuring Distance and Velocity with Defocus and Differential Motion

Best Paper Awards

Scene Chronology

Large-Scale Object Classification using Label Relation Graphs

Best Paper Awards – Honorable Mention

Visualizing and Understanding Convolutional Neural Networks

PAMI Everingham Prize

For extensive, generous, long-term service to the community in the management of computer vision conferences and workshops.

Koenderink Prize

High Accuracy Optical Flow Estimation 
Based on a Theory for Warping

Face Recognition with Local Binary Patterns

Best Paper Award

Segmentation Propagation in ImageNet

Best Paper Award – Honorable Mention

Activity Forecasting

Koenderink Prize

What Energy Functions Can Be Minimized via Graph Cuts?

Best Student Paper Award

Reconstructing the World’s Museums

Best Paper Award

Graph Cut based Inference with Co-occurrence Statistics

Runner-Up Paper Award

Blocks World Revisited: Image Understanding Using Qualitative Geometry and Mechanics

Koenderink Prize

Stochastic Tracking of 3D Human Figures Using 2D Image Motion (ECCV 2000)

For the ECCV’2000 paper entitled: Unsupervised Learning of Models for Recognition

Best Student Paper Award

Ambrosio-Tortorelli Segmentation of Stochastic Images

Best Paper Award

Learning Spatial Context: Using Stuff to Find Things

Koenderink Prize

Contour Tracking by Stochastic Propagation of Conditional Density (ECCV 1996)

Camera self-calibration: theory and experiments (ECCV 1992)

Best Student Paper Award

Learning to Localize Objects with Structured Output Regression

Longuet-Higgins Best Paper Award

Learning to Combine Bottom-up and Top-down Segmentation

Best Paper Award – Honorable Mention

Confocal Stereo

Simultaneous Object Pose and Velocity Computation Using a Single View from a Rolling Shutter Camera

Longuet-Higgins Best Paper Award

High Accuracy Optical Flow Estimation Based on a Theory for Warping

Best Paper Award – Honorable Mention

A Unified Algebraic Approach to 2-D and 3-D Motion Segmentation

Best Paper – Cognitive Vision

A Boosted Particle Filter: Multitarget Detection and Tracking

Iro Laina

Iro Laina’s dissertation addresses a broad range of challenging problems in computer vision, and specifically in scene understanding. The dissertation makes a series of timely, strong, and innovative scientific contributions by breaking down the overarching scene understanding problem to geometric, semantic, and linguistic components. It advances the state of the art in perceptual tasks and in the intersection of vision and language. The content of the thesis has been adopted in widespread real-world applications and has created significant impact.

Yongqin Xian

Yongqin Xian made significant contributions in the area of zero-shot and few-shot learning. He proposed novel feature generating frameworks that have defined the state of the art in this area. He introduced the first approach for zero-label semantic segmentation and pushed the boundary of few-shot video classification by leveraging unlabeled videos from a large dataset. In addition, his zero-shot learning benchmark has become highly influential in the field.

Marcella Cornia

Marcella Cornia is a cocktail of excellent technical expertise and great intellectual curiosity; the cherry on the glass is her kindness and willingness to work in both pure research projects, industrial research, and public engagement.

Triantafyllos Afouras

Triantafyllos Afouras’ outstanding dissertation introduces multiple creative ways to use audio and visual data in machine learning and in applications. It explores cross-modal and self-supervised learning for the important areas of lip reading, audio-visual speaker separation and enhancement, audio-visual object detection, and sign language recognition. Both the scientific outcomes of the dissertation and the datasets released have had a high impact in terms of citations and downloads.

Zeynep Akata

Zeynep Akata has had a number of significant contributions in multimodal deep learning, in vision and language for low-shot learning and in explainable machine learning. Her research defined the state of the art for zero-shot learning and her natural language explanations provided novel human-understandable justifications both for fine-grained visual classification and autonomous driving. She is a full professor at Tubingen University and leads a research group there. She has received a number of awards.

Amir Zamir

With a remarkable combination of creativity, productivity, and technical depth, Amir Zamir stands out as an emerging superstar in Computer Vision. Since earning his PhD, Prof. Zamir has pioneered highly influential approaches in multi-task/transfer learning and 3D environment simulation that serve as inspiration not simply for follow-up papers, but for entire research groups – both in academia and in industry. His work points the way forward as our field transitions from the era of “dataset AI” to the embodied AI future.