AIE-YOLO: Effective object detection method in extreme driving scenarios via adaptive image enhancement.

The widespread research and implementation of visual object detection technology have significantly transformed the autonomous driving industry. Autonomous driving relies heavily on visual sensors to perceive and analyze the environment. However, under extreme weather conditions, such as heavy rain, fog, or low light, these sensors may encounter disruptions, resulting in decreased image quality and reduced detection accuracy, thereby increasing the risk for autonomous driving. To address these challenges, we propose adaptive image enhancement (AIE)-YOLO, a novel object detection method to enhance road object detection accuracy under extreme weather conditions. To tackle the issue of image quality degradation in extreme weather, we designed an improved adaptive image enhancement module. This module dynamically adjusts the pixel features of road images based on different scene conditions, thereby enhancing object visibility and suppressing irrelevant background interference. Additionally, we introduce a spatial feature extraction module to adaptively enhance the model's spatial modeling capability under complex backgrounds. Furthermore, a channel feature extraction module is designed to adaptively enhance the model's representation and generalization abilities. Due to the difficulty in acquiring real-world data for various extreme weather conditions, we constructed a novel benchmark dataset named extreme weather simulation-rare object dataset. This dataset comprises ten types of simulated extreme weather scenarios and is built upon a publicly available rare object detection dataset. Extensive experiments conducted on the extreme weather simulation-rare object dataset demonstrate that AIE-YOLO outperforms existing state-of-the-art methods, achieving excellent detection performance under extreme weather conditions.

A lateral ankle sprain during a lateral backward step in badminton: A case report of a televised injury incident.

BACKGROUND: This study presents a kinematic analysis of an acute lateral ankle sprain incurred during a televised badminton match. The kinematics of this injury were compared to those of 19 previously reported cases in the published literature. METHODS: Four camera views of an acute lateral ankle sprain incurred during a televised badminton match were synchronized and rendered in 3-dimensional animation software. A badminton court with known dimensions was built in a virtual environment, and a skeletal model scaled to the injured athlete's height was used for skeletal matching. The ankle joint angle and angular velocity profiles of this acute injury were compared to the summarized findings from 19 previously reported cases in the published literature. RESULTS: At foot strike, the ankle joint was 2° everted, 33° plantarflexed, and 18° internally rotated. Maximum inversion of 114° and internal rotation of 69° was achieved at 0.24 s and 0.20 s after foot strike, respectively. After the foot strike, the ankle joint moved from an initial position of plantarflexion to dorsiflexion-from 33° plantarflexion to 53° dorsiflexion (range = 86°). Maximum inversion, dorsiflexion, and internal rotation angular velocity were 1262°/s, 961°/s, and 677°/s, respectively, at 0.12 s after foot strike. CONCLUSION: A forefoot landing posture with a plantarflexed and internally rotated ankle joint configuration could incite an acute lateral ankle sprain injury in badminton. Prevention of lateral ankle sprains in badminton should focus on the control and stability of the ankle joint angle during forefoot landings, especially when the athletes perform a combined lateral and backward step.

What have we learnt from quantitative case reports of acute lateral ankle sprains injuries and episodes of 'giving-way' of the ankle joint, and what shall we further investigate?

Lateral ankle sprains are a commonly incurred injury in sports. They have a high recurrence rate and can lead to the development of persistent injury associated symptoms. We performed a quantitative synthesis of published case reports documenting the kinematics of acute lateral ankle sprains and episodes of 'giving-way' of the ankle joint to provide a comprehensive description of the mechanisms. A systematic literature search was conducted to screen records within MEDLINE® and EMBASE®. Additional strategies included manual search of specific journals, as well as contacting researchers in relevant communities to retrieve unpublished data. Twenty-four cases were included in the quantitative synthesis, 11 from individual case reports and 13 from four separate case series. Two authors independently reviewed all the articles and extracted ankle joint kinematic data. Excessive ankle inversion was the most pronounced kinematic pattern observed across all included cases, with a mean peak inversion angle of 67.5° (range 2.0 to 142) and a mean peak inversion velocity of 974°/s (range 468 to 1752). This was followed by internal rotation and plantar flexion, respectively. A homogeneous linear function revealed a mean inversion velocity across all cases of 337°/s (range 117 to 1400; R2 = 0.78; p < 0.0001).

An Underwater Human-Robot Interaction Using a Visual-Textual Model for Autonomous Underwater Vehicles.

The marine environment presents a unique set of challenges for human-robot interaction. Communicating with gestures is a common way for interacting between the diver and autonomous underwater vehicles (AUVs). However, underwater gesture recognition is a challenging visual task for AUVs due to light refraction and wavelength color attenuation issues. Current gesture recognition methods classify the whole image directly or locate the hand position first and then classify the hand features. Among these purely visual approaches, textual information is largely ignored. This paper proposes a visual-textual model for underwater hand gesture recognition (VT-UHGR). The VT-UHGR model encodes the underwater diver's image as visual features, the category text as textual features, and generates visual-textual features through multimodal interactions. We guide AUVs to use image-text matching for learning and inference. The proposed method achieves better performance than most existing purely visual methods on the dataset CADDY, demonstrating the effectiveness of using textual patterns for underwater gesture recognition.

Whole mammographic mass segmentation using attention mechanism and multiscale pooling adversarial network.

Purpose: Since breast mass is a clear sign of breast cancer, its precise segmentation is of great significance for the diagnosis of breast cancer. However, the current diagnosis relies mainly on radiologists who spend time extracting features manually, which inevitably reduces the efficiency of diagnosis. Therefore, designing an automatic segmentation method is urgently necessary for the accurate segmentation of breast masses. Approach: We propose an effective attention mechanism and multiscale pooling conditional generative adversarial network (AM-MSP-cGAN), which accurately achieves mass automatic segmentation in whole mammograms. In AM-MSP-cGAN, U-Net is utilized as a generator network by incorporating attention mechanism (AM) into it, which allows U-Net to pay more attention to the target mass regions without additional cost. As a discriminator network, a convolutional neural network with multiscale pooling module is used to learn more meticulous features from the masses with rough and fuzzy boundaries. The proposed model is trained and tested on two public datasets: CBIS-DDSM and INbreast. Results: Compared with other state-of-the-art methods, the AM-MSP-cGAN can achieve better segmentation results in terms of the dice similarity coefficient (Dice) and Hausdorff distance metrics, achieving top scores of 84.49% and 5.01 on CBIS-DDSM, and 83.92% and 5.81 on INbreast, respectively. Therefore, qualitative and quantitative experiments illustrate that the proposed model is effective and robust for the mass segmentation in whole mammograms. Conclusions: The proposed deep learning model is suitable for the automatic segmentation of breast masses, which provides technical assistance for subsequent pathological structure analysis.

3D finger vein biometric authentication with photoacoustic tomography.

Biometric authentication is the recognition of human identity via unique anatomical features. The development of novel methods parallels widespread application by consumer devices, law enforcement, and access control. In particular, methods based on finger veins, as compared to face and fingerprints, obviate privacy concerns and degradation due to wear, age, and obscuration. However, they are two-dimensional (2D) and are fundamentally limited by conventional imaging and tissue-light scattering. In this work, for the first time, to the best of our knowledge, we demonstrate a method of three-dimensional (3D) finger vein biometric authentication based on photoacoustic tomography. Using a compact photoacoustic tomography setup and a novel recognition algorithm, the advantages of 3D are demonstrated via biometric authentication of index finger vessels with false acceptance, false rejection, and equal error rates <1.23%, <9.27%, and <0.13%, respectively, when comparing one finger, a false acceptance rateimprovement>10× when comparing multiple fingers, and <0.7% when rotating fingers ±30.

Dual Scan Mammoscope (DSM)-A New Portable Photoacoustic Breast Imaging System With Scanning in Craniocaudal Plane.

OBJECTIVE: We present a new photoacoustic tomography system that provides visualization of angiographic features in a human breast with mammogram-like images. METHODS: The system images a mildly compressed breast, from both top and bottom, using two 128-element, 2.25 MHz linear transducer arrays and line optical fiber bundles. The mild compression is achieved using plastic films, which is a more comfortable experience for the patient compared to rigid metal plates used in a traditional mammogram. RESULTS: We could image a D cup-sized breast of 7 cm thickness within 1 minute and achieve a spatial resolution of around 1 mm in all directions. CONCLUSION: Our system possesses the benefits of portability, speedy scanning, and patient comfort. The craniocaudal-view images can be easily correlated with existing imaging modalities for data interpretation. SIGNIFICANCE: Early cancer detection plays a critical role in overall cancer survival rate. Our system may address the limitations of mammogram and offer a radiation-free screening technique for patients with dense breasts.

A portable three-dimensional photoacoustic tomography system for imaging of chronic foot ulcers.

BACKGROUND: Chronic leg ulcers affect approximately 6.5 million Americans and the disorder is associated with a range of serious complications. Since many chronic ulcers have underlying vascular insufficiency, accurate assessment of tissue perfusion is critical to treatment planning and post-surgical monitoring. However, existing clinical tests fail to meet this need in practice due to their low sensitivity or accuracy. METHODS: In this paper, we introduce a portable photoacoustic tomography (PAT) system for wound assessment. Since hemoglobin serves as the major endogenous contrast at near-infrared wavelengths, PAT provides label-free, three-dimensional (3D) imaging of hemoglobin distribution, which is closely related to blood perfusion. The proposed system consists of a 128-element linear transducer array, a data acquisition (DAQ) system, and a pulsed Nd:YAG laser source, all mounted on a portable cart for easy clinical testing. RESULTS: We validated our system through both phantom and human imaging studies. The phantom imaging results indicate that the system's spatial resolution ranges from 0.5 mm along the axial direction to 1.3 mm along the elevational direction. The healthy volunteer result shows clear foot vasculature, indicating good perfusion. The preliminary patient imaging results agree very well with the clinical test, demonstrating that PAT has a high potential for assessing the circulation around the wound. CONCLUSIONS: We believe that our technique will be a valuable tool for assessing tissue perfusion and guiding wound treatment in vascular clinics.

Optimizing the light delivery of linear-array-based photoacoustic systems by double acoustic reflectors.

Although linear transducer arrays have been intensely used in photoacoustic imaging, their geometrical shape constrains light illumination. Today, most linear array based photoacoustic systems utilize side-illumination geometry, which consists of two line fiber bundles attached to the side of the probe. The angled light illumination increases the light travel distance in deep tissue, consequently limiting the imaging depth. This issue was partially addressed by adding a right angle prism in front of the transducer. While this design makes the light illumination and acoustic detection co-axial, the transducer and the fiber bundles are orthogonal to each other, making the system inconvenient for handheld use. To overcome this limitation, here we propose a double-reflector design, in which the second reflector redirects the acoustic signals by another 90°, so that the transducer and the fiber bundle are now parallel to each other. In this design, both the transducer and fiber bundle output are fitted into a compact housing for convenient handheld imaging. To evaluate the efficiency of our design, we performed various phantom and human in vivo experiments. Our results demonstrate that the double-reflector design indeed provides deeper imaging depth and it also allows for easy imaging of objects with uneven surfaces.

Optimizing photoacoustic image reconstruction using cross-platform parallel computation.

Three-dimensional (3D) image reconstruction involves the computations of an extensive amount of data that leads to tremendous processing time. Therefore, optimization is crucially needed to improve the performance and efficiency. With the widespread use of graphics processing units (GPU), parallel computing is transforming this arduous reconstruction process for numerous imaging modalities, and photoacoustic computed tomography (PACT) is not an exception. Existing works have investigated GPU-based optimization on photoacoustic microscopy (PAM) and PACT reconstruction using compute unified device architecture (CUDA) on either C++ or MATLAB only. However, our study is the first that uses cross-platform GPU computation. It maintains the simplicity of MATLAB, while improves the speed through CUDA/C++ - based MATLAB converted functions called MEXCUDA. Compared to a purely MATLAB with GPU approach, our cross-platform method improves the speed five times. Because MATLAB is widely used in PAM and PACT, this study will open up new avenues for photoacoustic image reconstruction and relevant real-time imaging applications.

Deep tissue photoacoustic computed tomography with a fast and compact laser system.

Photoacoustic computed tomography (PACT) holds great promise for biomedical imaging, but wide-spread implementation is impeded by the bulkiness of flash-lamp-pumped laser systems, which typically weigh between 50 - 200 kg, require continuous water cooling, and operate at a low repetition rate. Here, we demonstrate that compact lasers based on emerging diode technologies are well-suited for preclinical and clinical PACT. The diode-pumped laser used in this study had a miniature footprint (13 × 14 × 7 cm3), weighed only 1.6 kg, and outputted up to 80 mJ per pulse at 1064 nm. In vitro, the laser system readily provided over 4 cm PACT depth in chicken breast tissue. In vivo, in addition to high resolution, non-invasive brain imaging in living mice, the system can operate at 50 Hz, which enabled high-speed cross-sectional imaging of murine cardiac and respiratory function. The system also provided high quality, high-frame rate, and non-invasive three-dimensional mapping of arm, palm, and breast vasculature at multi centimeter depths in living human subjects, demonstrating the clinical viability of compact lasers for PACT.

Second generation slit-based photoacoustic tomography system for vascular imaging in human.

Slit-based photoacoustic tomography is a newly developed technique that improves the elevation numerical aperture of a linear array through acoustic diffraction. The slit, placed at the acoustic focus of a linear array, effectively forms an array of virtual detectors with high receiving angle, which subsequently improves the elevation resolution. However, due to the complex implementation, our original system could only image phantoms and sacrificed animals. In this report, the system has been significantly improved. In particular, we designed a slit holder that can be directly mounted to the transducer array for easy adjustment of slit width and simultaneous scanning of both the array and the slit. To enlarge the imaging field of view, we replaced the single circular optical fiber bundle with a bifurcated line fiber bundle which moved simultaneously with the array and the slit. The data acquisition system has also been updated to double the imaging speed. With these improvements, the new system can image a 3.8 × 4 cm2 region within 40 seconds and the object only needs to be coupled through ultrasound gel. We successfully used the system to image vasculatures in the palm and forearm of human volunteers. 3D palm vascular image of human palm.

Coherent-weighted three-dimensional image reconstruction in linear-array-based photoacoustic tomography.

While the majority of photoacoustic imaging systems used custom-made transducer arrays, commercially-available linear transducer arrays hold the benefits of affordable price, handheld convenience and wide clinical recognition. They are not widely used in photoacoustic imaging primarily because of the poor elevation resolution. Here, without modifying the imaging geometry and system, we propose addressing this limitation purely through image reconstruction. Our approach is based on the integration of two advanced image reconstruction techniques: focal-line-based three-dimensional image reconstruction and coherent weighting. We first numerically validated our approach through simulation and then experimentally tested it in phantom and in vivo. Both simulation and experimental results proved that the method can significantly improve the elevation resolution (up to 4 times in our experiment) and enhance object contrast.

A Phosphorus Phthalocyanine Formulation with Intense Absorbance at 1000 nm for Deep Optical Imaging.

Although photoacoustic computed tomography (PACT) operates with high spatial resolution in biological tissues deeper than other optical modalities, light scattering is a limiting factor. The use of longer near infrared wavelengths reduces scattering. Recently, the rational design of a stable phosphorus phthalocyanine (P-Pc) with a long wavelength absorption band beyond 1000 nm has been reported. Here, we show that when dissolved in liquid surfactants, P-Pc can give rise to formulations with absorbance of greater than 1000 (calculated for a 1 cm path length) at wavelengths beyond 1000 nm. Using the broadly accessible Nd:YAG pulse laser emission output of 1064 nm, P-Pc could be imaged through 11.6 cm of chicken breast with PACT. P-Pc accumulated passively in tumors following intravenous injection in mice as observed by PACT. Following oral administration, P-Pc passed through the intestine harmlessly, and PACT could be used to non-invasively observe intestine function. When the contrast agent placed under the arm of a healthy adult human, a PACT transducer on the top of the arm could readily detect P-Pc through the entire 5 cm limb. Thus, the approach of using contrast media with extreme absorption at 1064 nm readily enables high quality optical imaging in vitro and in vivo in humans at exceptional depths.

Slit-enabled linear-array photoacoustic tomography with near isotropic spatial resolution in three dimensions.

Due to its unique capability of visualizing optical absorption in deep tissues, photoacoustic tomography is increasingly used in biomedical imaging. Among various types of transducer arrays, the linear array is perhaps the most widely used in photoacoustic tomography because it is commercially available and readily allows ultrasound imaging. However, the three-dimensional imaging capability of a linear array is limited due to its poor elevational resolution. While various scanning schemes have been proposed to address this problem, they all suffer from long scanning time to the best of our knowledge. To address this issue, we introduce slit-enabled three-dimensional photoacoustic tomography. The metal slit, placed at the array focus, causes the incoming photoacoustic waves to diffract along the elevation direction and, hence, significantly improves the elevation detection aperture and resolution. We tested the new system in both phantoms and animals. The slit improves the elevation resolution by 10 times without compromising scanning time.