Faster R-CNN Method Research Articles

Object detection with class aware region proposal network and focused attention objective

In this paper, we propose a novel deep CNN-based framework to improve object detection performance. First, we introduce the Class Aware Region Proposal Network (CARPN) to produce high quality region proposals by using a new strategy for anchor generation, and by training the network with both bounding boxes and category labels of the objects. Instead of learning a binary object/non-object classifier for generating region proposals, we assign the class label to each anchor, and train the region proposal network with a multi-class loss. Second, we introduce the Focused Attention (FA) objective to encourage the network to learn features mainly from objects of interest while suppressing those features from the background region. As a result, false positive proposals caused by strong background features can be reduced to a large extent. Comprehensive experimental evaluations reveal that the proposed CARPN & FA framework remarkably outperforms the baseline Faster R-CNN method up to 4.1% mAP with a shallower network and 2.8% mAP with a deeper network, and achieves a better mAP than most of the latest state-of-the-art methods.

DEEP CROSS-DOMAIN BUILDING EXTRACTION FOR SELECTIVE DEPTH ESTIMATION FROM OBLIQUE AERIAL IMAGERY

Abstract. With the technological advancements of aerial imagery and accurate 3d reconstruction of urban environments, more and more attention has been paid to the automated analyses of urban areas. In our work, we examine two important aspects that allow online analysis of building structures in city models given oblique aerial image sequences, namely automatic building extraction with convolutional neural networks (CNNs) and selective real-time depth estimation from aerial imagery. We use transfer learning to train the Faster R-CNN method for real-time deep object detection, by combining a large ground-based dataset for urban scene understanding with a smaller number of images from an aerial dataset. We achieve an average precision (AP) of about 80 % for the task of building extraction on a selected evaluation dataset. Our evaluation focuses on both dataset-specific learning and transfer learning. Furthermore, we present an algorithm that allows for multi-view depth estimation from aerial image sequences in real-time. We adopt the semi-global matching (SGM) optimization strategy to preserve sharp edges at object boundaries. In combination with the Faster R-CNN, it allows a selective reconstruction of buildings, identified with regions of interest (RoIs), from oblique aerial imagery.

Detecting non-hardhat-use by a deep learning method from far-field surveillance videos

Abstract Hardhats are an important safety measure used to protect construction workers from accidents. However, accidents caused in ignorance of wearing hardhats still occur. In order to strengthen the supervision of construction workers to avoid accidents, automatic non-hardhat-use (NHU) detection technology can play an important role. Existing automatic methods of detecting hardhat avoidance are commonly limited to the detection of objects in near-field surveillance videos. This paper proposes the use of a high precision, high speed and widely applicable Faster R-CNN method to detect construction workers' NHU. To evaluate the performance of Faster R-CNN, more than 100,000 construction worker image frames were randomly selected from the far-field surveillance videos of 25 different construction sites over a period of more than a year. The research analyzed various visual conditions of the construction sites and classified image frames according to their visual conditions. The image frames were input into Faster R-CNN according to different visual categories. The experimental results demonstrate that the high precision, high recall and fast speed of the method can effectively detect construction workers' NHU in different construction site conditions, and can facilitate improved safety inspection and supervision.

Detection and diagnosis of colitis on computed tomography using deep convolutional neural networks.

Colitis refers to inflammation of the inner lining of the colon that is frequently associated with infection and allergic reactions. In this paper, we propose deep convolutional neural networks methods for lesion-level colitis detection and a support vector machine (SVM) classifier for patient-level colitis diagnosis on routine abdominal CT scans. The recently developed Faster Region-based Convolutional Neural Network (Faster RCNN) is utilized for lesion-level colitis detection. For each 2D slice, rectangular region proposals are generated by region proposal networks (RPN). Then, each region proposal is jointly classified and refined by a softmax classifier and bounding-box regressor. Two convolutional neural networks, eight layers of ZF net and 16 layers of VGG net are compared for colitis detection. Finally, for each patient, the detections on all 2D slices are collected and a SVM classifier is applied to develop a patient-level diagnosis. We trained and evaluated our method with 80 colitis patients and 80 normal cases using 4×4-fold cross validation. For lesion-level colitis detection, with ZF net, the mean of average precisions (mAP) were 48.7% and 50.9% for RCNN and Faster RCNN, respectively. The detection system achieved sensitivities of 51.4% and 54.0% at two false positives per patient for RCNN and Faster RCNN, respectively. With VGG net, Faster RCNN increased the mAP to 56.9% and increased the sensitivity to 58.4% at two false positive per patient. For patient-level colitis diagnosis, with ZF net, the average areas under the ROC curve (AUC) were 0.978±0.009 and 0.984±0.008 for RCNN and Faster RCNN method, respectively. The difference was not statistically significant with P=0.18. At the optimal operating point, the RCNN method correctly identified 90.4% (72.3/80) of the colitis patients and 94.0% (75.2/80) of normal cases. The sensitivity improved to 91.6% (73.3/80) and the specificity improved to 95.0% (76.0/80) for the Faster RCNN method. With VGG net, Faster RCNN increased the AUC to 0.986±0.007 and increased the diagnosis sensitivity to 93.7% (75.0/80) and specificity was unchanged at 95.0% (76.0/80). Colitis detection and diagnosis by deep convolutional neural networks is accurate and promising for future clinical application.

A model for fine-grained vehicle classification based on deep learning

A model for fine-grained vehicle classification based on deep learning is proposed to handle complicated transportation scene. This model comprises of two parts, vehicle detection model and vehicle fine-grained detection and classification model. Faster R-CNN method is adopted in vehicle detection model to extract single vehicle images from an image with clutter background which may contains serval vehicles. This step provides data for the next classification model. In vehicle fine-grained classification model, an image contains only one vehicle is fed into a CNN model to produce a feature, then a joint bayesian network is used to implement the fine-grained classification process. Experiments show that vehicle’s make and model can be recognized from transportation images effectively by using our method. Furthermore,in order to build a large scale database easier, this paper comes up with a novel network collaborative annotation mechanism.

Animal Detection From Highly Cluttered Natural Scenes Using Spatiotemporal Object Region Proposals and Patch Verification

In this paper, we consider the animal object detection and segmentation from wildlife monitoring videos captured by motion-triggered cameras, called camera-traps. For these types of videos, existing approaches often suffer from low detection rates due to low contrast between the foreground animals and the cluttered background, as well as high false positive rates due to the dynamic background. To address this issue, we first develop a new approach to generate animal object region proposals using multilevel graph cut in the spatiotemporal domain. We then develop a cross-frame temporal patch verification method to determine if these region proposals are true animals or background patches. We construct an efficient feature description for animal detection using joint deep learning and histogram of oriented gradient features encoded with Fisher vectors. Our extensive experimental results and performance comparisons over a diverse set of challenging camera-trap data demonstrate that the proposed spatiotemporal object proposal and patch verification framework outperforms the state-of-the-art methods, including the recent Faster-RCNN method, on animal object detection accuracy by up to 4.5%.