🔍 Research any topic with AI-powered citations — Try Researchly freeStart Researching
Home/Research/Give me papers with cnn projects
AI Research Answer

Give me papers with cnn projects

2 cited papers · April 2, 2026 · Powered by Researchly AI

🧠
TL;DR

The retrieved evidence covers several applied and foundational Convolutional Neural Network (CNN) projects spanning object detection, driver safety, saliency de…

The retrieved evidence covers several applied and foundational Convolutional Neural Network (CNN) projects spanning object detection, driver safety, saliency detection, model compression, and framework development.12

Ren et al. (2016) represents the most highly cited work in this set, with 52,964 citations.

1
Driver Safety Development Real Time Driver Drowsiness Detection System Based on Convolutional Neural NetworkMaryam Hashemi, Alireza Mirrashid et al.2020arXiv
View
2
Ristretto: A Framework for Empirical Study of Resource-Efficient Inference in Convolutional Neural NetworksPhilipp Gysel, Jon J. Pimentel et al.2018IEEE Transactions on Neural Networks and Learning Systems
View
  • Faster R-CNN (Region Proposal Network) — A CNN-based object detection framework that shares convolutional features between a Region Proposal Network and a detection network, achieving ~5 fps on GPU with state-of-the-art accuracy on PASCAL VOC and MS COCO. Ren et al. (2016)
  • Caffe Framework — A BSD-licensed C++ deep learning library with Python/MATLAB bindings, capable of processing over 40 million images per day on a single GPU. Jia et al. (2014)
  • Driver Drowsiness Detection CNN — A real-time system using CNNs (including Transfer Learning on VGG16/VGG19) for eye-closure-based drowsiness detection, with a newly proposed eye dataset.
1Hashemi et al. (2020)1
  • Visual Saliency Detection — A multiscale CNN-based approach that uses deep contrast features to achieve state-of-the-art saliency detection performance. Li & Yu (2016)
  • Resource-Efficient CNN Inference — The Ristretto framework empirically studies word-width approximations for deploying CNNs on resource-constrained embedded systems.
2Gysel et al. (2018)2
1
Driver Safety Development Real Time Driver Drowsiness Detection System Based on Convolutional Neural NetworkMaryam Hashemi, Alireza Mirrashid et al.2020arXiv
View
2
Ristretto: A Framework for Empirical Study of Resource-Efficient Inference in Convolutional Neural NetworksPhilipp Gysel, Jon J. Pimentel et al.2018IEEE Transactions on Neural Networks and Learning Systems
View
Want to research your own topic? Try it free →
Diagram
CNN-Based Project Landscape (from retrieved papers)
─────────────────────────────────────────────────────
Input Image
 │
 ▼
[Feature Extraction — CNN Backbone]
 │
 ├──► Object Detection (Faster R-CNN)
 │ └─ RPN → Region Proposals → Fast R-CNN Head
 │
 ├──► Driver Safety (Drowsiness CNN)
 │ └─ Eye Region → FD-NN / TL-VGG16/VGG19 → Sleep State
 │
 ├──► Visual Saliency (Multiscale CNN)
 │ └─ 3-Scale Features → Deep Contrast Feature → Saliency Map
 │
 └──► Model Compression (Ristretto)
 └─ Pretrained CNN → Approximation → Embedded Deployment
Table
ProjectCNN ArchitectureKey Result / Benchmark
Faster R-CNNVGG-16 + RPN5 fps GPU; SOTA on PASCAL VOC 2007/2012 & MS COCO
Drowsiness DetectionVGG16/VGG19 (Transfer Learning) + custom FD-NNHigh accuracy, low computational complexity on eye closure dataset
Visual SaliencyMultiscale deep CNNF-measure improvement of 6.12%–10% over prior SOTA on public benchmarks
RistrettoGeneric CNN approximationEmpirically studies word-width vs. accuracy tradeoff for embedded deployment
Want to research your own topic? Try it free →
  • The drowsiness detection system notes a significant lack of available and accurate eye datasets, which the authors attempt to address with a newly proposed dataset, but broader generalizability remains unverified.
1
  • Ristretto's approximation framework introduces potential accuracy degradation when reducing word width, and the tradeoff between compression and classification accuracy requires empirical evaluation.
2
1
Driver Safety Development Real Time Driver Drowsiness Detection System Based on Convolutional Neural NetworkMaryam Hashemi, Alireza Mirrashid et al.2020arXiv
View
2
Ristretto: A Framework for Empirical Study of Resource-Efficient Inference in Convolutional Neural NetworksPhilipp Gysel, Jon J. Pimentel et al.2018IEEE Transactions on Neural Networks and Learning Systems
View
  • Faster R-CNN introduced the Region Proposal Network to eliminate the region proposal bottleneck, achieving real-time-class detection at 5 fps on GPU.
1
  • Caffe enables processing of over 40 million images per day on a single GPU, making it suitable for industrial-scale CNN deployment.
  • Multiscale CNN features improve visual saliency detection by 6.12%–10% in F-measure over prior state-of-the-art methods.
  • Transfer learning on VGG16/VGG19 provides a viable path for real-time driver drowsiness detection with high accuracy and low computational cost.
1
1
Driver Safety Development Real Time Driver Drowsiness Detection System Based on Convolutional Neural NetworkMaryam Hashemi, Alireza Mirrashid et al.2020arXiv
View
Want to research your own topic? Try it free →
  1. "CNN-based real-time object detection benchmarks PASCAL VOC MS COCO 2022–2024"
  2. "Transfer learning VGG ResNet for medical image classification Indian datasets"
  3. "Model compression and efficient inference for embedded systems"

Research smarter with AI-powered citations

Researchly finds and cites academic papers for any research topic in seconds. Used by students across India.

Try Free See Pricing