Hybridizing Convolutional Neural Network for Classification of Lung diseases

Pulmonary disease is widespread worldwide. There is persistent blockage of the lungs, pneumonia, asthma, TB, etc. It is essential to diagnose the lungs promptly. For this reason, machine learning models were developed. For lung disease prediction, many deep learning technologies, including the CNN, and the capsule network, are used. The fundamental CNN has low rotating, inclined, or other irregular image orientation efficiency. Therefore by integrating the space transformer network (STN) with CNN, we propose a new hybrid deep learning architecture named STNCNN. The new model is implemented on the dataset from the Kaggle repository for an NIH chest X-ray image. STNCNN has an accuracy of 69% in respect of the entire dataset, while the accuracy values of vanilla grey, vanilla RGB, hybrid CNN are 67.8%, 69.5%, and 63.8%, respectively. When the sample data set is applied, STNCNN takes much less time to train at the cost of a slightly less reliable validation. Therefore both specialists and physicians are simplified by the proposed STNCNN System for the diagnosis of lung disease.

Comparison of Gradient-Based Edge Detectors Applied on Mammograms

Breast cancer is one of the most common types of cancer amongst women, but it is also one of the most frequently cured cancers. Because of this, early detection is crucial, and this can be done through mammography screening. With the increasing need of an automated interpretation system, a lot of methods have been proposed so far and, regardless of the algorithms, they all share a step: pre-processing. That is, identifying the image orientation, detecting the breast and eliminating irrelevant parts. This paper aims to describe, analyze, compare and evaluate six of the most commonly used edge detection operators: Sobel, Roberts Cross, Prewitt, Farid and Simoncelli, Scharr and Canny. We detail the algorithms, their implementations and the metrics used for evaluation and continue by comparing the operators both visually and numerically, finally concluding that Canny best suit our needs.

Archivists and Architects: A Case Study on Mitigating Design Cataloging Challenges through Copyright Outreach

ABSTRACT For many archives working to publish collections online, securing copyright is a time-consuming challenge. What if the labor-intensive process of copyright outreach could be designed to increase the yield of staff time and add value to collection metadata? This case study explores an effort at the Irwin S. Chanin School of Architecture Archive of The Cooper Union to combine copyright outreach with author-generated metadata in an attempt to address common architectural record cataloging challenges. Specifically, the study looks at direct correspondence with creators as an opportunity to both secure permission to publish copyrighted materials and to fill descriptive holes in subject, title, caption, image orientation, and authorized name. This article further studies the feasibility of combining two laborintensive processes, discusses a preliminary and revised workflow, and evaluates the practicality and value of corresponding with over eight hundred individuals.

Enhanced Magnetic Resonance Image Synthesis with Contrast-Aware Generative Adversarial Networks

A magnetic resonance imaging (MRI) exam typically consists of the acquisition of multiple MR pulse sequences, which are required for a reliable diagnosis. With the rise of generative deep learning models, approaches for the synthesis of MR images are developed to either synthesize additional MR contrasts, generate synthetic data, or augment existing data for AI training. While current generative approaches allow only the synthesis of specific sets of MR contrasts, we developed a method to generate synthetic MR images with adjustable image contrast. Therefore, we trained a generative adversarial network (GAN) with a separate auxiliary classifier (AC) network to generate synthetic MR knee images conditioned on various acquisition parameters (repetition time, echo time, and image orientation). The AC determined the repetition time with a mean absolute error (MAE) of 239.6 ms, the echo time with an MAE of 1.6 ms, and the image orientation with an accuracy of 100%. Therefore, it can properly condition the generator network during training. Moreover, in a visual Turing test, two experts mislabeled 40.5% of real and synthetic MR images, demonstrating that the image quality of the generated synthetic and real MR images is comparable. This work can support radiologists and technologists during the parameterization of MR sequences by previewing the yielded MR contrast, can serve as a valuable tool for radiology training, and can be used for customized data generation to support AI training.

A Robust Hand Silhouette Orientation Detection Method for Hand Gesture Recognition

The computer vision approach is most widely used for research related to hand gesture recognition. The detection of the image orientation has been discovered to be one of the keys to determine its success. The degree of freedom for a hand determines the shape and orientation of a gesture, which further causes a problem in the recognition methods. This article proposed evaluating orientation detection for silhouette static hand gestures with different poses and orientations without considering the forearm. The longest chord and ellipse were the two popular methods compared. The angles formed from two wrist points were selected as ground truth data and calculated from the horizontal axis. The performance was analyzed using the error values obtained from the difference in ground truth data angles compared to the method's results. The method has errors closer to zero that were rated better. Moreover, the method was evaluated using 1187 images, divided into four groups based on the forearm presence, and the results showed its effect on orientation detection. It was also discovered that the ellipse method was better than the longest chord. This study's results are used to select hand gesture orientation detection to increase accuracy in the hand gesture recognition process.

EVALUATING HAND-CRAFTED AND LEARNING-BASED FEATURES FOR PHOTOGRAMMETRIC APPLICATIONS

The image orientation (or Structure from Motion – SfM) process needs well localized, repeatable and stable tie points in order to derive camera poses and a sparse 3D representation of the surveyed scene. The accurate identification of tie points in large image datasets is still an open research topic in the photogrammetric and computer vision communities. Tie points are established by firstly extracting keypoint using a hand-crafted feature detector and descriptor methods. In the last years new solutions, based on convolutional neural network (CNN) methods, were proposed to let a deep network discover which feature extraction process and representation are most suitable for the processed images. In this paper we aim to compare state-of-the-art hand-crafted and learning-based method for the establishment of tie points in various and different image datasets. The investigation highlights the actual challenges for feature matching and evaluates selected methods under different acquisition conditions (network configurations, image overlap, UAV vs terrestrial, strip vs convergent) and scene's characteristics. Remarks and lessons learned constrained to the used datasets and methods are provided.

PREFACE: TECHNICAL COMMISSION II

ISPRS Technical Commission II focuses, at various scales, on geometric, radiometric and multi-temporal aspects of image- and range-based 3D surveying and modeling. Specifically, Commission II deals with image orientation, point cloud generation and processing, 3D feature extraction, dynamic and static scene analysis, sensor and data fusion, and machine learning for geospatial data analysis and big data techniques for massive data processing. Applications in the fields of mapping, infrastructure monitoring, heritage studies, space exploration, underwater photogrammetry and environmental engineering are also considered.The Volume related to Commission II contains 108 papers published in the ISPRS Archives and 20 published in the ISPRS Annals. The papers in Archives were accepted from among 111 abstract submissions and 59 full papers. Of the 108 Archives papers, 32 were full paper submissions and 76 were abstract submissions. The 20 papers in the Annals were selected from among 59 full paper submissions after going through a peer review process.In the aforementioned areas of research, the papers in this volume discuss the challenges and needs, and introduce novel photogrammetric solutions that depict the latest developments in the field.There has been a wide range of coverage of these topics and point cloud generation and processing has been the most active coverage with close to 38% of the accepted papers. This is followed by machine/deep learning methods with 31.8% that provide solutions to the semantic enrichment of images and 3D data. Research in heritage and underwater studies is represented with 9% and 6%, respectively, of the accepted papers.We believe these volumes nicely recap the state-of-the-art, current trends and possible applications of photogrammetry on a wide range of topics with a nice overview of the future research directions.On behalf of Technical Commission II, we would like to thank the local organizers of the 2021 ISPRS Congress, the members of the international program committee, all Working Group officers and all reviewers for their hard organizational work and efforts in the paper reviewing process.

EVALUATING TIE POINTS DISTRIBUTION, MULTIPLICITY AND NUMBER ON THE ACCURACY OF UAV PHOTOGRAMMETRY BLOCKS

Image orientation is a fundamental task in photogrammetric applications and it is performed by extracting keypoints with hand-crafted or learning-based methods, generating tie points among the images and running a bundle adjustment procedure. Nowadays, due to large number of extracted keypoints, tie point filtering approaches attempt to eliminate redundant tie points in order to increase accuracy and reduce processing time. This paper presents the results of an investigation concerning tie points impact on bundle adjustment results. Simulations and real data are processed in Australis and DBAT to evaluate different affecting factors, including tie point numbers, location accuracy, distribution and multiplicity. Achieved results show that increasing the amount of tie points improve the quality of bundle adjustment results, provided that the tie points are well-distributed on the image. Furthermore, bundle adjustment quality is improved as the multiplicity of tie points increases and their location uncertainty decrease. Based on simulation results, some suggestions for accurate tie points filtering in typical UAV photogrammetry blocks cases are derived.

Imaging Attitude Control and Image Motion Compensation Residual Analysis Based on a Three-Axis Inertially Stabilized Platform

The airborne area camera has received broad application in aerial reconnaissance, land resource surveying, environmental monitoring, photogrammetry mapping, and natural disaster information acquisition. A three-axis, inertially stabilized platform with a large rotation range for the roll axis is designed, which is based on the cantilever structure, in order to realize a large-angle sweep imaging function for airborne area cameras. An image attitude control algorithm in the inertial space is proposed, which can regulate the line of sight (LOS) as well as the image orientation. The area camera image motion calculation model and image motion compensation residual computing method are proposed, utilizing space position and velocity vector transformation mathematics and derivations. The variation of linear velocity of the image motion in the sensor frame is analyzed, and the changing laws of the maximum deviation of image motion with the image attitude are studied. Flight tests imply that the vertical imaging technique correctly regulates the LOS along the local geodetic vertical. The along-flight overlap rate is greater than 65%, which meets the stereo mapping requirement. The sweep imaging technique considerably enlarges the cross-flight angle of view. The LOS and image orientation during sweep imaging are correctly controlled, and gap-free coverage of the survey area is maintained. The image’s azimuth or roll deviation is less than 0.1°, and the image pitch deviation is less than 0.35°. The quality of the test images is superior. Black and white line pairs for evaluation can be clearly distinguished. The image’s motion is well compensated, and the image motion compensation residual is well constrained. These verify the validity of the proposed imaging technique and the image motion analysis model.