Results 71 to 80 of about 86,091 (310)

Coherentice: Invertible Concept-Based Explainability Framework for CNNs beyond Fidelity

In their natural form, convolutional neural networks (CNNs) lack interpretability despite their effectiveness in visual categorization. Concept activation vectors (CAVs) offer human-interpretable quantitative explainability, utilizing feature maps from ...
Gao, Y, Zhou, J, Akpudo, UE, Lewis, A
core   +1 more source

Improving the Robustness of Visual Teach‐and‐Repeat Navigation Using Drift Error Correction and Event‐Based Vision for Low‐Light Environments

Advanced Robotics Research, EarlyView.
Visual teach‐and‐repeat (VTR) navigation allows robots to learn and follow routes without building a full metric map. We show that navigation accuracy for VTR can be improved by integrating a topological map with error‐drift correction based on stereo vision.
Fuhai Ling, Ze Huang, Tony J. Prescott
wiley   +1 more source

IJCM_21A: How do under-five children belonging to scheduled tribes compare with the non- tribal children: a secondary data analysis using the Comprehensive National Nutritional Survey (CNNS) 2016-18.

Indian Journal of Community Medicine
Background: Scheduled tribes (ST) constitute 8.6% of India’s population. Disproportionate burden of undernutrition is found among these socially disadvantaged population.
H Pavithra, Madhavi Bhargava, Konsam Dinachandra   +2 more
doaj   +1 more source

TraNCE: Transformative Nonlinear Concept Explainer for CNNs

Convolutional neural networks (CNNs) have succeeded remarkably in various computer vision tasks. However, they are not intrinsically explainable. While feature-level understanding of CNNs reveals where the models looked, concept-based explainability ...
Akpudo, Ugochukwu Ejike, Lewis, Andrew, Zhou, Jun, Gao, Yongsheng   +3 more
core   +1 more source

Comparison of CNNs and SVM for voice control wheelchair

, 2020
In this paper, we develop an intelligent wheelchair using CNNs and SVM voice recognition methods. The data is collected from Google and some of them are self-recorded.
Mohd Ali, Azliza, Sharifuddin, Mohammad Shahrul Izham, Mohammad Shahrul Izham Sharifuddin, Sharifalillah Nordin, Azliza Mohd Ali, Nordin, Sharifalillah   +5 more
core   +1 more source

Multimodal Human–Robot Interaction Using Human Pose Estimation and Local Large Language Models

Advanced Robotics Research, EarlyView.
A multimodal human–robot interaction framework integrates human pose estimation (HPE) and a large language model (LLM) for gesture‐ and voice‐based robot control. Speech‐to‐text (STT) enables voice command interpretation, while a safety‐aware arbitration mechanism prioritizes gesture input for rapid intervention.
Nasiru Aboki, Ilche Georgievski, Marco Aiello   +2 more
wiley   +1 more source

How explainable are adversarially-robust CNNs?

, 2023
Three important criteria of existing convolutional neural networks (CNNs) are (1) test-set accuracy; (2) out-of-distribution accuracy; and (3) explainability. While these criteria have been studied independently, their relationship is unknown.
Chen, Peijie, Nguyen, Anh, Nourelahi, Mehdi, Kotthoff, Lars   +3 more
core  

Data‐Driven Bulldozer Blade Control for Autonomous Terrain Leveling

Advanced Robotics Research, EarlyView.
A simulation‐driven framework for autonomous bulldozer leveling is presented, combining high‐fidelity terramechanics simulation with a neural‐network‐based reduced‐order model. Gradient‐based optimization enables efficient, low‐level blade control that balances leveling quality and operation time.
Harry Zhang, Ganesh Arivoli, Huzaifa Unjhawala, Luning Bakke, Radu Serban, Dan Negrut   +5 more
wiley   +1 more source

Combining bag of visual words-based features with CNN in image classification

Journal of Intelligent Systems
Although traditional image classification techniques are often used in authentic ways, they have several drawbacks, such as unsatisfactory results, poor classification accuracy, and a lack of flexibility.
Marzouk Marwa A., Elkholy Mohamed
doaj   +1 more source

LAND USE CLASSIFICATION BASED ON MULTI-STRUCTURE CONVOLUTION NEURAL NETWORK FEATURES CASCADING [PDF]

The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2019
Learning efficient image representations is at the core of the classification task of remote sensing imagery. The existing methods for solving image classification task, based on either feature coding approaches extracted from convolution neural networks(
J. Men, L. Fang, Y. Liu, Y. Sun
doaj   +1 more source