Search Results for author: Lihi Zelnik-Manor

Found 30 papers, 20 papers with code

Diverse Imagenet Models Transfer Better

no code implementations19 Apr 2022 Niv Nayman, Avram Golbert, Asaf Noy, Tan Ping, Lihi Zelnik-Manor

Encouraged by the recent transferability results of self-supervised models, we propose a method that combines self-supervised and supervised pretraining to generate models with both high diversity and high accuracy, and as a result high transferability.

Multi-Label Classification

BINAS: Bilinear Interpretable Neural Architecture Search

1 code implementation24 Oct 2021 Niv Nayman, Yonathan Aflalo, Asaf Noy, Rong Jin, Lihi Zelnik-Manor

Practical use of neural networks often involves requirements on latency, energy and memory among others.

Neural Architecture Search

Multi-label Classification with Partial Annotations using Class-aware Selective Loss

1 code implementation CVPR 2022 Emanuel Ben-Baruch, Tal Ridnik, Itamar Friedman, Avi Ben-Cohen, Nadav Zamir, Asaf Noy, Lihi Zelnik-Manor

We propose to estimate the class distribution using a dedicated temporary model, and we show its improved efficiency over a naive estimation computed using the dataset's partial annotations.

Multi-Label Classification

PETA: Photo Albums Event Recognition using Transformers Attention

1 code implementation26 Sep 2021 Tamar Glaser, Emanuel Ben-Baruch, Gilad Sharir, Nadav Zamir, Asaf Noy, Lihi Zelnik-Manor

We address this gap with a tailor-made solution, combining the power of CNNs for image representation and transformers for album representation to perform global reasoning on image collection, offering a practical and efficient solution for photo albums event recognition.

Semantic Diversity Learning for Zero-Shot Multi-label Classification

1 code implementation ICCV 2021 Avi Ben-Cohen, Nadav Zamir, Emanuel Ben Baruch, Itamar Friedman, Lihi Zelnik-Manor

We argue that using a single embedding vector to represent an image, as commonly practiced, is not sufficient to rank both relevant seen and unseen labels accurately.

Classification Image Retrieval +3

An Image is Worth 16x16 Words, What is a Video Worth?

2 code implementations25 Mar 2021 Gilad Sharir, Asaf Noy, Lihi Zelnik-Manor

Methods that reach State of the Art (SotA) accuracy, usually make use of 3D convolution layers as a way to abstract the temporal information from video frames.

Ranked #17 on Action Recognition on UCF101 (using extra training data)

Action Classification Action Recognition

HardCoRe-NAS: Hard Constrained diffeRentiable Neural Architecture Search

2 code implementations23 Feb 2021 Niv Nayman, Yonathan Aflalo, Asaf Noy, Lihi Zelnik-Manor

Realistic use of neural networks often requires adhering to multiple constraints on latency, energy and memory among others.

Neural Architecture Search

A Convergence Theory Towards Practical Over-parameterized Deep Neural Networks

no code implementations12 Jan 2021 Asaf Noy, Yi Xu, Yonathan Aflalo, Lihi Zelnik-Manor, Rong Jin

We show that convergence to a global minimum is guaranteed for networks with widths quadratic in the sample size and linear in their depth at a time logarithmic in both.

Compact Network Training for Person ReID

no code implementations15 Oct 2019 Hussam Lawen, Avi Ben-Cohen, Matan Protter, Itamar Friedman, Lihi Zelnik-Manor

Furthermore, we show the representation power of our ReID network via SotA results on a different task of multi-object tracking.

Ranked #12 on Person Re-Identification on Market-1501 (Rank-1 metric)

Multi-Object Tracking Person Re-Identification

XNAS: Neural Architecture Search with Expert Advice

2 code implementations NeurIPS 2019 Niv Nayman, Asaf Noy, Tal Ridnik, Itamar Friedman, Rong Jin, Lihi Zelnik-Manor

This paper introduces a novel optimization method for differential neural architecture search, based on the theory of prediction with expert advice.

Image Classification Neural Architecture Search

ASAP: Architecture Search, Anneal and Prune

1 code implementation8 Apr 2019 Asaf Noy, Niv Nayman, Tal Ridnik, Nadav Zamir, Sivan Doveh, Itamar Friedman, Raja Giryes, Lihi Zelnik-Manor

In this paper, we propose a differentiable search space that allows the annealing of architecture weights, while gradually pruning inferior operations.

Neural Architecture Search

Is Image Memorability Prediction Solved?

no code implementations31 Jan 2019 Shay Perera, Ayellet Tal, Lihi Zelnik-Manor

This paper deals with the prediction of the memorability of a given image.

Dynamic-Net: Tuning the Objective Without Re-training for Synthesis Tasks

1 code implementation ICCV 2019 Alon Shoshan, Roey Mechrez, Lihi Zelnik-Manor

Our approach considers an "objective-space" as the space of all linear combinations of two objectives, and the Dynamic-Net is emulating the traversing of this objective-space at test-time, without any further training.

Adversarial Feedback Loop

1 code implementation ICCV 2019 Firas Shama, Roey Mechrez, Alon Shoshan, Lihi Zelnik-Manor

In this paper we propose a novel method that makes an explicit use of the discriminator in test-time, in a feedback manner in order to improve the generator results.

Image Generation Super-Resolution

The 2018 PIRM Challenge on Perceptual Image Super-resolution

6 code implementations20 Sep 2018 Yochai Blau, Roey Mechrez, Radu Timofte, Tomer Michaeli, Lihi Zelnik-Manor

This paper reports on the 2018 PIRM challenge on perceptual super-resolution (SR), held in conjunction with the Perceptual Image Restoration and Manipulation (PIRM) workshop at ECCV 2018.

Image Restoration Image Super-Resolution

Modifying Non-Local Variations Across Multiple Views

no code implementations CVPR 2018 Tal Tlusty, Tomer Michaeli, Tali Dekel, Lihi Zelnik-Manor

We present an algorithm for modifying small non-local variations between repeating structures and patterns in multiple images of the same scene.

Maintaining Natural Image Statistics with the Contextual Loss

3 code implementations13 Mar 2018 Roey Mechrez, Itamar Talmi, Firas Shama, Lihi Zelnik-Manor

Maintaining natural image statistics is a crucial factor in restoration and generation of realistic looking images.

Image Super-Resolution Single Image Super Resolution +1

The Contextual Loss for Image Transformation with Non-Aligned Data

3 code implementations ECCV 2018 Roey Mechrez, Itamar Talmi, Lihi Zelnik-Manor

Feed-forward CNNs trained for image transformation problems rely on loss functions that measure the similarity between the generated image and a target image.

Photorealistic Style Transfer with Screened Poisson Equation

1 code implementation28 Sep 2017 Roey Mechrez, Eli Shechtman, Lihi Zelnik-Manor

Recent work has shown impressive success in transferring painterly style to images.

Style Transfer

Saliency Driven Image Manipulation

1 code implementation7 Dec 2016 Roey Mechrez, Eli Shechtman, Lihi Zelnik-Manor

Have you ever taken a picture only to find out that an unimportant background object ended up being overly salient?

Image Manipulation

Template Matching with Deformable Diversity Similarity

1 code implementation CVPR 2017 Itamar Talmi, Roey Mechrez, Lihi Zelnik-Manor

We propose a novel measure for template matching named Deformable Diversity Similarity -- based on the diversity of feature matches between a target image window and the template.

Template Matching

Hot or Not: Exploring Correlations Between Appearance and Temperature

no code implementations ICCV 2015 Daniel Glasner, Pascal Fua, Todd Zickler, Lihi Zelnik-Manor

In this paper we explore interactions between the appearance of an outdoor scene and the ambient temperature.

Approximate Nearest Neighbor Fields in Video

no code implementations CVPR 2015 Nir Ben-Zrihem, Lihi Zelnik-Manor

We introduce RIANN (Ring Intersection Approximate Nearest Neighbor search), an algorithm for matching patches of a video to a set of reference patches in real-time.

Colorization Denoising

How to Evaluate Foreground Maps?

no code implementations CVPR 2014 Ran Margolin, Lihi Zelnik-Manor, Ayellet Tal

In this paper, we show that the most commonly-used measures for evaluating both non-binary maps and binary maps do not always provide a reliable evaluation.

object-detection RGB Salient Object Detection +2

What Makes a Patch Distinct?

no code implementations CVPR 2013 Ran Margolin, Ayellet Tal, Lihi Zelnik-Manor

Our key contribution is a novel and fast approach to compute pattern distinctness.

Learning Video Saliency from Human Gaze Using Candidate Selection

no code implementations CVPR 2013 Dmitry Rudoy, Dan B. Goldman, Eli Shechtman, Lihi Zelnik-Manor

For example, the time each video frame is observed is a fraction of a second, while a still image can be viewed leisurely.

Saliency Prediction

Crowdsourcing Gaze Data Collection

1 code implementation16 Apr 2012 Dmitry Rudoy, Dan B. Goldman, Eli Shechtman, Lihi Zelnik-Manor

In this work we propose a crowdsourced method for acquisition of gaze direction data from a virtually unlimited number of participants, using a robust self-reporting mechanism (see Figure 1).

Social and Information Networks Human-Computer Interaction

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