Document Layout Analysis Using Detection Transformers

2021 ◽  
Author(s):  
Prashanth Pillai ◽  
Purnaprajna Mangsuli
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Superior Performance ◽  
Layout Analysis ◽  
Bounding Box ◽  
Document Layout Analysis ◽  
Wide Range ◽  
Document Layout ◽  
Bounding Boxes ◽  
Document Page

Abstract In the O&G (Oil & Gas) industry, unstructured data sources such as technical reports on hydrocarbon production, daily drilling, well construction, etc. contain valuable information. This information however is conveyed through various formats such as tables, forms, text, figures, etc. Detecting these different entities in documents is essential for building a structured representation of the information within and for automated processing of documents at scale. Our work presents a document layout analysis workflow to detect/localize different entities based on a deep learning-based framework. The workflow comprises of a deep learning-based object-detection framework based on transformers to identify the spatial location of entities in a document page. The key elements of the object-detection pipeline include a residual network backbone for feature extraction and an encoder-decoder transformer based on the latest detection transformers (DETR) to predict object-bounding boxes and category labels. The object detection is formulated as a direct set prediction task using bipartite matching while also eliminating conventional operations like anchor box generation and non-maximal suppression. The availability of sufficient publicly available document layout data sets that incorporate the artifacts observed in historical O&G technical reports is often a major challenge. We attempt to address this challenge by using a novel training data augmentation methodology. The dense occurrence of elements in a page can often introduce uncertainties resulting in bounding boxes cutting through text content. We adopt a bounding box post-processing methodology to refine the bounding box coordinates to minimize undercuts. The proposed document layout analysis pipeline was trained to detect entity types such as headings, text blocks, tables, forms, and images/charts in a document page. A wide range of pages from lithology, stratigraphy, drilling, and field development reports were used for model training. The reports also included a considerable number of historical scanned reports. The trained object-detection model was evaluated on a test data set prepared from the O&G reports. DETR demonstrated superior performance when compared with the Mask R-CNN on our dataset.

scholarly journals Pano-RSOD: A Dataset and Benchmark for Panoramic Road Scene Object Detection

Electronics ◽  
2019 ◽  
Vol 8 (3) ◽  
pp. 329 ◽  
Author(s):  
Yong Li ◽  
Guofeng Tong ◽  
Huashuai Gao ◽  
Yuebin Wang ◽  
Liqiang Zhang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Training Data ◽  
Training Dataset ◽  
Panoramic Image ◽  
Information Object ◽  
Wide Range ◽  
Panoramic Images ◽  
Bounding Boxes ◽  
Scene Object

Panoramic images have a wide range of applications in many fields with their ability to perceive all-round information. Object detection based on panoramic images has certain advantages in terms of environment perception due to the characteristics of panoramic images, e.g., lager perspective. In recent years, deep learning methods have achieved remarkable results in image classification and object detection. Their performance depends on the large amount of training data. Therefore, a good training dataset is a prerequisite for the methods to achieve better recognition results. Then, we construct a benchmark named Pano-RSOD for panoramic road scene object detection. Pano-RSOD contains vehicles, pedestrians, traffic signs and guiding arrows. The objects of Pano-RSOD are labelled by bounding boxes in the images. Different from traditional object detection datasets, Pano-RSOD contains more objects in a panoramic image, and the high-resolution images have 360-degree environmental perception, more annotations, more small objects and diverse road scenes. The state-of-the-art deep learning algorithms are trained on Pano-RSOD for object detection, which demonstrates that Pano-RSOD is a useful benchmark, and it provides a better panoramic image training dataset for object detection tasks, especially for small and deformed objects.

Iterative Scheme for Object Detection in Crowded Environments

2021 ◽  
Vol 12 (1) ◽  
pp. 31-39
Author(s):  
D. D. Rukhovich ◽  
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Iterative Scheme ◽  
State Of The Art ◽  
Sufficient Accuracy ◽  
Two Stage ◽  
One Stage ◽  
Bounding Box ◽  
Bounding Boxes ◽  
Crowded Environments

Deep learning-based detectors usually produce a redundant set of object bounding boxes including many duplicate detections of the same object. These boxes are then filtered using non-maximum suppression (NMS) in order to select exactly one bounding box per object of interest. This greedy scheme is simple and provides sufficient accuracy for isolated objects but often fails in crowded environments, since one needs to both preserve boxes for different objects and suppress duplicate detections. In this work we develop an alternative iterative scheme, where a new subset of objects is detected at each iteration. Detected boxes from the previous iterations are passed to the network at the following iterations to ensure that the same object would not be detected twice. This iterative scheme can be applied to both one-stage and two-stage object detectors with just minor modifications of the training and inference proce­dures. We perform extensive experiments with two different baseline detectors on four datasets and show significant improvement over the baseline, leading to state-of-the-art performance on CrowdHuman and WiderPerson datasets.

scholarly journals Improved Oriented Object Detection in Remote Sensing Images Based on a Three-Point Regression Method

2021 ◽  
Vol 13 (22) ◽  
pp. 4517
Author(s):  
Falin Wu ◽  
Jiaqi He ◽  
Guopeng Zhou ◽  
Haolun Li ◽  
Yushuang Liu ◽  
...  
Keyword(s):  
Remote Sensing ◽  
Object Detection ◽  
Poor Performance ◽  
Regression Method ◽  
Remote Sensing Images ◽  
Sensing Applications ◽  
Bounding Box ◽  
Bounding Boxes ◽  
Fully Connected ◽  
Oriented Object

Object detection in remote sensing images plays an important role in both military and civilian remote sensing applications. Objects in remote sensing images are different from those in natural images. They have the characteristics of scale diversity, arbitrary directivity, and dense arrangement, which causes difficulties in object detection. For objects with a large aspect ratio and that are oblique and densely arranged, using an oriented bounding box can help to avoid deleting some correct detection bounding boxes by mistake. The classic rotational region convolutional neural network (R2CNN) has advantages for text detection. However, R2CNN has poor performance in the detection of slender objects with arbitrary directivity in remote sensing images, and its fault tolerance rate is low. In order to solve this problem, this paper proposes an improved R2CNN based on a double detection head structure and a three-point regression method, namely, TPR-R2CNN. The proposed network modifies the original R2CNN network structure by applying a double fully connected (2-fc) detection head and classification fusion. One detection head is for classification and horizontal bounding box regression, the other is for classification and oriented bounding box regression. The three-point regression method (TPR) is proposed for oriented bounding box regression, which determines the positions of the oriented bounding box by regressing the coordinates of the center point and the first two vertices. The proposed network was validated on the DOTA-v1.5 and HRSC2016 datasets, and it achieved a mean average precision (mAP) of 3.90% and 15.27%, respectively, from feature pyramid network (FPN) baselines with a ResNet-50 backbone.

scholarly journals A Survey on Various Available Object Detection Models and Application In Automatic License Plate Detection

2021 ◽  
Vol 23 (06) ◽  
pp. 47-57
Author(s):  
Aditya Kulkarni ◽  
◽  
Manali Munot ◽  
Sai Salunkhe ◽  
Shubham Mhaske ◽  
...  
Keyword(s):  
Deep Learning ◽  
Object Detection ◽  
Image Databases ◽  
License Plate ◽  
Learning Models ◽  
Python Language ◽  
Performance Accuracy ◽  
License Plate Detection ◽  
Bounding Boxes ◽  
Complex Images

With the development in technologies right from serial to parallel computing, GPU, AI, and deep learning models a series of tools to process complex images have been developed. The main focus of this research is to compare various algorithms(pre-trained models) and their contributions to process complex images in terms of performance, accuracy, time, and their limitations. The pre-trained models we are using are CNN, R-CNN, R-FCN, and YOLO. These models are python language-based and use libraries like TensorFlow, OpenCV, and free image databases (Microsoft COCO and PAS-CAL VOC 2007/2012). These not only aim at object detection but also on building bounding boxes around appropriate locations. Thus, by this review, we get a better vision of these models and their performance and a good idea of which models are ideal for various situations.

scholarly journals EFN: Field-Based Object Detection for Aerial Images

2020 ◽  
Vol 12 (21) ◽  
pp. 3630
Author(s):  
Jin Liu ◽  
Haokun Zheng
Keyword(s):  
Object Detection ◽  
Semantic Segmentation ◽  
Natural Images ◽  
Aerial Images ◽  
Aerial Image ◽  
Data Sets ◽  
Data Set ◽  
Bounding Box ◽  
Good Score ◽  
Bounding Boxes

Object detection and recognition in aerial and remote sensing images has become a hot topic in the field of computer vision in recent years. As these images are usually taken from a bird’s-eye view, the targets often have different shapes and are densely arranged. Therefore, using an oriented bounding box to mark the target is a mainstream choice. However, this general method is designed based on horizontal box annotation, while the improved method for detecting an oriented bounding box has a high computational complexity. In this paper, we propose a method called ellipse field network (EFN) to organically integrate semantic segmentation and object detection. It predicts the probability distribution of the target and obtains accurate oriented bounding boxes through a post-processing step. We tested our method on the HRSC2016 and DOTA data sets, achieving mAP values of 0.863 and 0.701, respectively. At the same time, we also tested the performance of EFN on natural images and obtained a mAP of 84.7 in the VOC2012 data set. These extensive experiments demonstrate that EFN can achieve state-of-the-art results in aerial image tests and can obtain a good score when considering natural images.

Sign in / Sign up

Export Citation Format

Share Document