Search Results for author: Ali Harakeh

Found 14 papers, 7 papers with code

Self-Supervised Image-to-Point Distillation via Semantically Tolerant Contrastive Loss

no code implementations CVPR 2023 Anas Mahmoud, Jordan S. K. Hu, Tianshu Kuai, Ali Harakeh, Liam Paull, Steven L. Waslander

However, image-to point representation learning for autonomous driving datasets faces two main challenges: 1) the abundance of self-similarity, which results in the contrastive losses pushing away semantically similar point and image regions and thus disturbing the local semantic structure of the learned representations, and 2) severe class imbalance as pretraining gets dominated by over-represented classes.

3D Semantic Segmentation Autonomous Driving +4

Estimating Regression Predictive Distributions with Sample Networks

no code implementations24 Nov 2022 Ali Harakeh, Jordan Hu, Naiqing Guan, Steven L. Waslander, Liam Paull

A common approach to model uncertainty is to choose a parametric distribution and fit the data to it using maximum likelihood estimation.

regression

Bayesian Embeddings for Few-Shot Open World Recognition

no code implementations29 Jul 2021 John Willes, James Harrison, Ali Harakeh, Chelsea Finn, Marco Pavone, Steven Waslander

As autonomous decision-making agents move from narrow operating environments to unstructured worlds, learning systems must move from a closed-world formulation to an open-world and few-shot setting in which agents continuously learn new classes from small amounts of information.

Decision Making Few-Shot Learning

Estimating and Evaluating Regression Predictive Uncertainty in Deep Object Detectors

3 code implementations13 Jan 2021 Ali Harakeh, Steven L. Waslander

We show that in the context of object detection, training variance networks with negative log likelihood (NLL) can lead to high entropy predictive distributions regardless of the correctness of the output mean.

Object object-detection +3

Predictive Uncertainty in Deep Object Detectors: Estimation and Evaluation

no code implementations ICLR 2021 Ali Harakeh, Steven L. Waslander

We show that in the context of object detection, training variance networks with negative log likelihood (NLL) can lead to high entropy predictive distributions regardless of the correctness of the output mean.

Object object-detection +2

A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving

1 code implementation20 Nov 2020 Di Feng, Ali Harakeh, Steven Waslander, Klaus Dietmayer

Next, we present a strict comparative study for probabilistic object detection based on an image detector and three public autonomous driving datasets.

Autonomous Driving Object +2

BayesOD: A Bayesian Approach for Uncertainty Estimation in Deep Object Detectors

2 code implementations9 Mar 2019 Ali Harakeh, Michael Smart, Steven L. Waslander

When incorporating deep neural networks into robotic systems, a major challenge is the lack of uncertainty measures associated with their output predictions.

Object object-detection +1

Leveraging Pre-Trained 3D Object Detection Models For Fast Ground Truth Generation

no code implementations16 Jul 2018 Jungwook Lee, Sean Walsh, Ali Harakeh, Steven L. Waslander

Training 3D object detectors for autonomous driving has been limited to small datasets due to the effort required to generate annotations.

3D Object Detection Autonomous Driving +6

Unlimited Road-scene Synthetic Annotation (URSA) Dataset

no code implementations16 Jul 2018 Matt Angus, Mohamed ElBalkini, Samin Khan, Ali Harakeh, Oles Andrienko, Cody Reading, Steven Waslander, Krzysztof Czarnecki

Utilizing open-source tools and resources found in single-player modding communities, we provide a method for persistent, ground truth, asset annotation of a game world.

Semantic Segmentation

A Hierarchical Deep Architecture and Mini-Batch Selection Method For Joint Traffic Sign and Light Detection

1 code implementation20 Jun 2018 Alex D. Pon, Oles Andrienko, Ali Harakeh, Steven L. Waslander

The root cause of this issue is that no public dataset contains both traffic light and sign labels, which leads to difficulties in developing a joint detection framework.

Traffic Sign Detection Traffic Sign Recognition

In Defense of Classical Image Processing: Fast Depth Completion on the CPU

2 code implementations31 Jan 2018 Jason Ku, Ali Harakeh, Steven L. Waslander

With the rise of data driven deep neural networks as a realization of universal function approximators, most research on computer vision problems has moved away from hand crafted classical image processing algorithms.

Depth Completion Test

Identifying Good Training Data for Self-Supervised Free Space Estimation

no code implementations CVPR 2016 Ali Harakeh, Daniel Asmar, Elie Shammas

This paper proposes a novel technique to extract training data from free space in a scene using a stereo camera.

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