Homography-guided stereo matching for wide-baseline image interpolation

Image interpolation has a wide range of applications such as frame rate-up conversion and free viewpoint TV. Despite significant progresses, it remains an open challenge especially for image pairs with large displacements. In this paper, we first propose a novel optimization algorithm for motion estimation, which combines the advantages of both global optimization and a local parametric transformation model. We perform optimization over dynamic label sets, which are modified after each iteration using the prior of piecewise consistency to avoid local minima. Then we apply it to an image interpolation framework including occlusion handling and intermediate image interpolation. We validate the performance of our algorithm experimentally, and show that our approach achieves state-of-the-art performance.

Review of Wide-Baseline Stereo Image Matching Based on Deep Learning

Strong geometric and radiometric distortions often exist in optical wide-baseline stereo images, and some local regions can include surface discontinuities and occlusions. Digital photogrammetry and computer vision researchers have focused on automatic matching for such images. Deep convolutional neural networks, which can express high-level features and their correlation, have received increasing attention for the task of wide-baseline image matching, and learning-based methods have the potential to surpass methods based on handcrafted features. Therefore, we focus on the dynamic study of wide-baseline image matching and review the main approaches of learning-based feature detection, description, and end-to-end image matching. Moreover, we summarize the current representative research using stepwise inspection and dissection. We present the results of comprehensive experiments on actual wide-baseline stereo images, which we use to contrast and discuss the advantages and disadvantages of several state-of-the-art deep-learning algorithms. Finally, we conclude with a description of the state-of-the-art methods and forecast developing trends with unresolved challenges, providing a guide for future work.

Qualitative Visual Assessment of Hand Hygiene Product Effectiveness

Background: Effective hand hygiene (HH) is an essential preventative measure for the reduction of hospital-acquired infections (HAIs). Commonly used HH products include alcohol-based hand rubs (ABHRs), antimicrobial soaps, and nonantimicrobial soaps. In vivo clinical studies have demonstrated that levels of bacterial reduction can vary based on the HH product type, formulation, and dose. It has been reported that ABHRs provide the greatest reduction in bacteria, followed by antimicrobial soaps. Objective: We examined the effects of products representative of 3 HH categories on artificially soiled hands, using a hand-stamp procedure. The hand-stamp images provide a clear visualization of product effectiveness and can be used as an educational tool to promote the importance of proper hand hygiene using different product formats. Method: Three commercially available formulations were evaluated in this study, a mild nonantimicrobial soap, an antimicrobial soap containing chloroxylenol (PCMX), and an ABHR containing 70% v/v ethanol. Prior to the hand stamp procedure, the participant’s hands were prewashed with 5 mL of a nonantimicrobial soap and dried. An inoculum of Serratia marcescens containing ∼1 × 109CFU/mL was prepared as described in ASTM E2755. A 0.2-mL aliquot of the inoculum was dispensed onto the palm of the subject’s hand and spread by rubbing over the entire surface of both hands. Following a 30-second dry time, one of the subject’s hands was gently pressed onto the surface of a large petri dish containing tryptic soy agar to obtain a baseline image. Following the baseline sample, 1 pump of the selected test product (∼0.9 mL for soap or 1.1 mL for ABHR) was applied to the participant’s hands. For soap applications, hands were vigorously rubbed for 30 seconds followed by a 30-second water rinse. For ABHR, product was rubbed by the user until dry. The hand-stamp procedure was repeated following product application using the participant’s other hand. Results: Clear qualitative reductions in bacteria were observed with each of the HH interventions. The greatest reduction was observed following the application of ABHR. Antimicrobial soap was less effective than ABHR but more effective than nonantimicrobial soap. Conclusions: The qualitative visual model demonstrates the effectiveness of various HH interventions and correlates with log reductions observed in traditional efficacy test methods. Future efforts should explore hand-stamp repeatability and image utilization to support HH improvement efforts in healthcare systems.Funding: GOJO Industries provided support for this study.Disclosures: Mary Rose Czaplicki reports salary from GOJO Industries.

Prediction of lesion shrinkage using CT imaging with radiomic and deep learning approaches.

e14592 Background: Radiomic features derived from CT scans have shown promise in predicting treatment response (Sun et al 2018, and others). We carried out a proof-of-concept study to investigate the use of CT images to predict lesion-level response. Methods: CT images from Merck studies KEYNOTE-010 (NCT01905657) and KEYNOTE-024 (NCT02142738), were used. Data from each study were evaluated separately and split for training (80%) and validation (20%) in each study. A lesion was classified as “shrinking” if ≥30% size reduction from baseline was seen on any future scan. There were 2004 (613 shrinking vs. 1391 non-shrinking) and 588 (311 vs. 277) lesions in KN10 and KN24, respectively. 130 radiomic features were extracted, followed by random forest to predict lesion response. In addition, end-to-end deep learning was used, which predicts the response directly from ROIs of CT images. Models were trained in two ways: (1) using pre-treatment baseline (BL) only or (2) using both BL and the first post-treatment image (V1) as predictors. Finally, to evaluate the predictive power without relying on initial lesion size, size information was omitted from CT images. Results: Results from the KN10 and KN24 are summarized in Table. Conclusions: The results suggest that the BL CT images alone have little power to predict lesion response, while BL and the first post-baseline image exhibit high predictive power. Although a substantial part of the predictive power can be attributed to change in ROI size, the predictive power does exist in other aspects of CT images. Overall, the radiomic signature followed by random forest produced predictions similar to, if not better than, the deep learning approach. [Table: see text]

NETWORK DESIGN IN CLOSE-RANGE PHOTOGRAMMETRY WITH SHORT BASELINE IMAGES

The avaibility of automated software for image-based 3D modelling has changed the way people acquire images for photogrammetric applications. Short baseline images are required to match image points with SIFT-like algorithms, obtaining more images than those necessary for “old fashioned” photogrammetric projects based on manual measurements. This paper describes some considerations on network design for short baseline image sequences, especially on precision and reliability of bundle adjustment. Simulated results reveal that the large number of 3D points used for image orientation has very limited impact on network precision.