Search Results for author: Krystian Mikolajczyk

Found 43 papers, 13 papers with code

Language-Based Depth Hints for Monocular Depth Estimation

no code implementations22 Mar 2024 Dylan Auty, Krystian Mikolajczyk

In this work, we demonstrate the use of natural language as a source of an explicit prior about the structure of the world.

Instance Segmentation Language Modelling +2

Learning to Project for Cross-Task Knowledge Distillation

no code implementations21 Mar 2024 Dylan Auty, Roy Miles, Benedikt Kolbeinsson, Krystian Mikolajczyk

In this setting, cross-task distillation can be used, enabling the use of any teacher model trained on a different task.

Depth Estimation Knowledge Distillation +2

Match-Stereo-Videos: Bidirectional Alignment for Consistent Dynamic Stereo Matching

no code implementations16 Mar 2024 Junpeng Jing, Ye Mao, Krystian Mikolajczyk

Towards this challenge, we develop a bidirectional alignment mechanism for adjacent frames as a fundamental operation.

Stereo Matching

Adaptive Early Exiting for Collaborative Inference over Noisy Wireless Channels

no code implementations29 Nov 2023 Mikolaj Jankowski, Deniz Gunduz, Krystian Mikolajczyk

In this work, we study early exiting in the context of collaborative inference, which allows obtaining inference results at the edge device for certain samples, without the need to transmit the partially processed data to the edge server at all, leading to further communication savings.

Collaborative Inference Image Classification

Features-over-the-Air: Contrastive Learning Enabled Cooperative Edge Inference

no code implementations17 Apr 2023 Haotian Wu, Nitish Mital, Krystian Mikolajczyk, Deniz Gündüz

We study the collaborative image retrieval problem at the wireless edge, where multiple edge devices capture images of the same object, which are then used jointly to retrieve similar images at the edge server over a shared multiple access channel.

Contrastive Learning Image Retrieval +1

Understanding the Role of the Projector in Knowledge Distillation

4 code implementations20 Mar 2023 Roy Miles, Krystian Mikolajczyk

We then show that the normalisation of representations is tightly coupled with the training dynamics of this projector, which can have a large impact on the students performance.

Image Classification Knowledge Distillation +3

Multi-Class Segmentation from Aerial Views using Recursive Noise Diffusion

1 code implementation1 Dec 2022 Benedikt Kolbeinsson, Krystian Mikolajczyk

Semantic segmentation from aerial views is a crucial task for autonomous drones, as they rely on precise and accurate segmentation to navigate safely and efficiently.

Denoising Navigate +2

ObjCAViT: Improving Monocular Depth Estimation Using Natural Language Models And Image-Object Cross-Attention

1 code implementation30 Nov 2022 Dylan Auty, Krystian Mikolajczyk

While monocular depth estimation (MDE) is an important problem in computer vision, it is difficult due to the ambiguity that results from the compression of a 3D scene into only 2 dimensions.

Image-to-Image Translation Monocular Depth Estimation +1

SAMPLE-HD: Simultaneous Action and Motion Planning Learning Environment

no code implementations1 Jun 2022 Michal Nazarczuk, Tony Ng, Krystian Mikolajczyk

Humans exhibit incredibly high levels of multi-modal understanding - combining visual cues with read, or heard knowledge comes easy to us and allows for very accurate interaction with the surrounding environment.

Motion Planning Question Answering +2

Monocular Depth Estimation Using Cues Inspired by Biological Vision Systems

1 code implementation21 Apr 2022 Dylan Auty, Krystian Mikolajczyk

Part of the MDE task is, therefore, to learn which visual cues in the image can be used for depth estimation, and how.

Depth Prediction Instance Segmentation +2

NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning

no code implementations CVPR 2022 Tony Ng, Hyo Jin Kim, Vincent Lee, Daniel DeTone, Tsun-Yi Yang, Tianwei Shen, Eddy Ilg, Vassileios Balntas, Krystian Mikolajczyk, Chris Sweeney

We let a feature encoding network and image reconstruction network compete with each other, such that the feature encoder tries to impede the image reconstruction with its generated descriptors, while the reconstructor tries to recover the input image from the descriptors.

Camera Localization Image Reconstruction

Information Theoretic Representation Distillation

1 code implementation1 Dec 2021 Roy Miles, Adrian Lopez Rodriguez, Krystian Mikolajczyk

Despite the empirical success of knowledge distillation, current state-of-the-art methods are computationally expensive to train, which makes them difficult to adopt in practice.

Classification with Binary Weight Network Knowledge Distillation +1

Reconstructing Pruned Filters using Cheap Spatial Transformations

no code implementations25 Oct 2021 Roy Miles, Krystian Mikolajczyk

We present an efficient alternative to the convolutional layer using cheap spatial transformations.

Feature Compression Knowledge Distillation +1

Grasp-Oriented Fine-grained Cloth Segmentation without Real Supervision

no code implementations6 Oct 2021 Ruijie Ren, Mohit Gurnani Rajesh, Jordi Sanchez-Riera, Fan Zhang, Yurun Tian, Antonio Agudo, Yiannis Demiris, Krystian Mikolajczyk, Francesc Moreno-Noguer

We show that training our network solely with synthetic data and the proposed DA yields results competitive with models trained on real data.

Domain Adaptation

AirNet: Neural Network Transmission over the Air

no code implementations24 May 2021 Mikolaj Jankowski, Deniz Gunduz, Krystian Mikolajczyk

We further improve the performance of AirNet by pruning the network below the available bandwidth, and expanding it for improved robustness.

Knowledge Distillation

DESC: Domain Adaptation for Depth Estimation via Semantic Consistency

no code implementations3 Sep 2020 Adrian Lopez-Rodriguez, Krystian Mikolajczyk

Instead, in this paper, we propose a domain adaptation approach to train a monocular depth estimation model using a fully-annotated source dataset and a non-annotated target dataset.

Domain Adaptation Monocular Depth Estimation +1

Cascaded channel pruning using hierarchical self-distillation

no code implementations16 Aug 2020 Roy Miles, Krystian Mikolajczyk

In this paper, we propose an approach for filter-level pruning with hierarchical knowledge distillation based on the teacher, teaching-assistant, and student framework.

Knowledge Distillation Model Compression

Wireless Image Retrieval at the Edge

no code implementations21 Jul 2020 Mikolaj Jankowski, Deniz Gunduz, Krystian Mikolajczyk

We propose two alternative schemes based on digital and analog communications, respectively.

Image Compression Image Retrieval +1

HyNet: Learning Local Descriptor with Hybrid Similarity Measure and Triplet Loss

1 code implementation NeurIPS 2020 Yurun Tian, Axel Barroso-Laguna, Tony Ng, Vassileios Balntas, Krystian Mikolajczyk

Recent works show that local descriptor learning benefits from the use of L2 normalisation, however, an in-depth analysis of this effect lacks in the literature.

3D Reconstruction Patch Matching +1

D2D: Keypoint Extraction with Describe to Detect Approach

no code implementations27 May 2020 Yurun Tian, Vassileios Balntas, Tony Ng, Axel Barroso-Laguna, Yiannis Demiris, Krystian Mikolajczyk

In this paper, we present a novel approach that exploits the information within the descriptor space to propose keypoint locations.

3D Reconstruction

HDD-Net: Hybrid Detector Descriptor with Mutual Interactive Learning

1 code implementation12 May 2020 Axel Barroso-Laguna, Yannick Verdie, Benjamin Busam, Krystian Mikolajczyk

Local feature extraction remains an active research area due to the advances in fields such as SLAM, 3D reconstructions, or AR applications.

3D Reconstruction

SHOP-VRB: A Visual Reasoning Benchmark for Object Perception

no code implementations6 Apr 2020 Michal Nazarczuk, Krystian Mikolajczyk

In this paper we present an approach and a benchmark for visual reasoning in robotics applications, in particular small object grasping and manipulation.

Object Visual Reasoning

Domain Adversarial Training for Infrared-colour Person Re-Identification

no code implementations9 Mar 2020 Nima Mohammadi Meshky, Sara Iodice, Krystian Mikolajczyk

However, in poorly-lit environments CCTV cameras switch to infrared imaging, hence developing a system which can correctly perform matching between infrared and colour images is a necessity.

Person Re-Identification

Joint Device-Edge Inference over Wireless Links with Pruning

no code implementations4 Mar 2020 Mikolaj Jankowski, Deniz Gunduz, Krystian Mikolajczyk

We propose a joint feature compression and transmission scheme for efficient inference at the wireless network edge.

Feature Compression General Classification +2

SOLAR: Second-Order Loss and Attention for Image Retrieval

1 code implementation ECCV 2020 Tony Ng, Vassileios Balntas, Yurun Tian, Krystian Mikolajczyk

One is focused on second-order spatial information to increase the performance of image descriptors, both local and global.

Image Retrieval Patch Matching +1

Compression of descriptor models for mobile applications

no code implementations9 Jan 2020 Roy Miles, Krystian Mikolajczyk

Deep neural networks have demonstrated state-of-the-art performance for feature-based image matching through the advent of new large and diverse datasets.

Domain Adaptation for Object Detection via Style Consistency

1 code implementation22 Nov 2019 Adrian Lopez Rodriguez, Krystian Mikolajczyk

For the first step, we use a style transfer method for pixel-adaptation of source images to the target domain.

Domain Adaptation Object +3

Deep Joint Source-Channel Coding for Wireless Image Retrieval

no code implementations28 Oct 2019 Mikolaj Jankowski, Deniz Gunduz, Krystian Mikolajczyk

Motivated by surveillance applications with wireless cameras or drones, we consider the problem of image retrieval over a wireless channel.

Image Retrieval Retrieval

Key.Net: Keypoint Detection by Handcrafted and Learned CNN Filters

3 code implementations ICCV 2019 Axel Barroso-Laguna, Edgar Riba, Daniel Ponsa, Krystian Mikolajczyk

We introduce a novel approach for keypoint detection task that combines handcrafted and learned CNN filters within a shallow multi-scale architecture.

Ranked #5 on Image Matching on IMC PhotoTourism (using extra training data)

Image Matching Keypoint Detection

Person Re-identification with Bias-controlled Adversarial Training

no code implementations30 Mar 2019 Sara Iodice, Krystian Mikolajczyk

Inspired by the effectiveness of adversarial training in the area of Generative Adversarial Networks we present a new approach for learning feature representations in person re-identification.

Person Re-Identification

Partial Person Re-identification with Alignment and Hallucination

no code implementations24 Jul 2018 Sara Iodice, Krystian Mikolajczyk

We propose Partial Matching Net (PMN) that detects body joints, aligns partial views and hallucinates the missing parts based on the information present in the frame and a learned model of a person.

Hallucination Person Re-Identification

Deep Segmentation and Registration in X-Ray Angiography Video

1 code implementation16 May 2018 Athanasios Vlontzos, Krystian Mikolajczyk

In interventional radiology, short video sequences of vein structure in motion are captured in order to help medical personnel identify vascular issues or plan intervention.

Optical Flow Estimation Position +2

Person Re-Identification with Vision and Language

no code implementations3 Oct 2017 Fei Yan, Krystian Mikolajczyk, Josef Kittler

We propose a joint vision and language model based on CCA and CNN architectures to match across the two modalities as well as to enrich visual examples for which there are no language descriptions.

Language Modelling Person Re-Identification

The BreakingNews Dataset

no code implementations WS 2017 Arnau Ramisa, Fei Yan, Francesc Moreno-Noguer, Krystian Mikolajczyk

We present BreakingNews, a novel dataset with approximately 100K news articles including images, text and captions, and enriched with heterogeneous meta-data (e. g. GPS coordinates and popularity metrics).

Image Captioning Sentiment Analysis

BreakingNews: Article Annotation by Image and Text Processing

no code implementations23 Mar 2016 Arnau Ramisa, Fei Yan, Francesc Moreno-Noguer, Krystian Mikolajczyk

Deep Canonical Correlation Analysis is deployed for article illustration, and a new loss function based on Great Circle Distance is proposed for geolocation.

Image Retrieval Retrieval +1

PN-Net: Conjoined Triple Deep Network for Learning Local Image Descriptors

1 code implementation19 Jan 2016 Vassileios Balntas, Edward Johns, Lilian Tang, Krystian Mikolajczyk

We address this problem and propose a CNN based descriptor with improved matching performance, significantly reduced training and execution time, as well as low dimensionality.

Deep Correlation for Matching Images and Text

no code implementations CVPR 2015 Fei Yan, Krystian Mikolajczyk

This paper addresses the problem of matching images and captions in a joint latent space learnt with deep canonical correlation analysis (DCCA).

BOLD - Binary Online Learned Descriptor For Efficient Image Matching

no code implementations CVPR 2015 Vassileios Balntas, Lilian Tang, Krystian Mikolajczyk

The patch adapted descriptors are then efficiently built online from a subset of tests which lead to lower intra class distances thus a more robust descriptor.

Cannot find the paper you are looking for? You can Submit a new open access paper.