Automatically Detect and Track Multiple Fish Swimming in Shallow Water with Frequent Occlusion

Due to its universality, swarm behavior in nature attracts much attention of scientists from many fields. Fish schools are examples of biological communities that demonstrate swarm behavior. The detection and tracking of fish in a school are of important significance for the quantitative research on swarm behavior. However, different from other biological communities, there are three problems in the detection and tracking of fish school, that is, variable appearances, complex motion and frequent occlusion. To solve these problems, we propose an effective method of fish detection and tracking. In this method, first, the fish head region is positioned through extremum detection and ellipse fitting; second, The Kalman filtering and feature matching are used to track the target in complex motion; finally, according to the feature information obtained by the detection and tracking, the tracking problems caused by frequent occlusion are processed through trajectory linking. We apply this method to track swimming fish school of different densities. The experimental results show that the proposed method is both accurate and reliable.     

A Deep-Structured Conditional Random Field Model for Object Silhouette Tracking

In this work, we introduce a deep-structured conditional random field (DS-CRF) model for the purpose of state-based object silhouette tracking. The proposed DS-CRF model consists of a series of state layers, where each state layer spatially characterizes the object silhouette at a particular point in time. The interactions between adjacent state layers are established by inter-layer connectivity dynamically determined based on inter-frame optical flow. By incorporate both spatial and temporal context in a dynamic fashion within such a deep-structured probabilistic graphical model, the proposed DS-CRF model allows us to develop a framework that can accurately and efficiently track object silhouettes that can change greatly over time, as well as under different situations such as occlusion and multiple targets within the scene. Experiment results using video surveillance datasets containing different scenarios such as occlusion and multiple targets showed that the proposed DS-CRF approach provides strong object silhouette tracking performance when compared to baseline methods such as mean-shift tracking, as well as state-of-the-art methods such as context tracking and boosted particle filtering.     

How lesioning the nucleus praeeminentialis affects electrolocation behavior in the weakly electric fish, Apteronotus leptorhynchus

The ability to orient to and track moving electrolocation targets was assessed in normal Apteronotus leptorhynchus and in those with unilateral lesions of the nucleus praeeminentialis dorsalis.

1. Each fish was trained to hover between two vertical metal rods and track their movement. Two aspects of this behavior were measured: a) the hovering position of the fish relative to stationary rods; b) the latency between the onset of rod motion and the fish's tracking response. Control fish hovered midway between stationary rods, while lesioned fish hovered closer to the rod ipsilateral to the lesion. Response latency varied negatively with rod diameter in both sets of fish, and lesioned fish exhibited shorter latencies than control fish. While the response latencies of control fish were shortest when their starting position was midway between the rods, lesioned animals' latencies were shortest when they hovered closer to the rod ipsilateral to their lesion.
2. Control fish responded to the approach of a single metal ball to either side of the body with nearly equal latencies and fish-to-object distances. After lesioning, response latency increased and fish-to-object distance decreased for approaches to the side ipsilateral to the lesion; opposite changes occurred for contralateral approaches.

     

Uncovering the Mechanism of Trapping and Cell Orientation during Neisseria gonorrhoeae Twitching Motility

Neisseria gonorrheae bacteria are the causative agent of the second most common sexually transmitted infection in the world. The bacteria move on a surface by means of twitching motility. Their movement is mediated by multiple long and flexible filaments, called type IV pili, that extend from the cell body, attach to the surface, and retract, thus generating a pulling force. Moving cells also use pili to aggregate and form microcolonies. However, the mechanism by which the pili surrounding the cell body work together to propel bacteria remains unclear. Understanding this process will help describe the motility of N. gonorrheae bacteria, and thus the dissemination of the disease which they cause. In this article we track individual twitching cells and observe that their trajectories consist of alternating moving and pausing intervals, while the cell body is preferably oriented with its wide side toward the direction of motion. Based on these data, we propose a model for the collective pili operation of N. gonorrheae bacteria that explains the experimentally observed behavior. Individual pili function independently but can lead to coordinated motion or pausing via the force balance. The geometry of the cell defines its orientation during motion. We show that by changing pili substrate interactions, the motility pattern can be altered in a predictable way. Although the model proposed is tangibly simple, it still has sufficient robustness to incorporate further advanced pili features and various cell geometries to describe other bacteria that employ pili to move on surfaces.     

Uncovering the Mechanism of Trapping and Cell Orientation during Neisseria gonorrhoeae Twitching Motility

Neisseria gonorrheae bacteria are the causative agent of the second most common sexually transmitted infection in the world. The bacteria move on a surface by means of twitching motility. Their movement is mediated by multiple long and flexible filaments, called type IV pili, that extend from the cell body, attach to the surface, and retract, thus generating a pulling force. Moving cells also use pili to aggregate and form microcolonies. However, the mechanism by which the pili surrounding the cell body work together to propel bacteria remains unclear. Understanding this process will help describe the motility of N. gonorrheae bacteria, and thus the dissemination of the disease which they cause. In this article we track individual twitching cells and observe that their trajectories consist of alternating moving and pausing intervals, while the cell body is preferably oriented with its wide side toward the direction of motion. Based on these data, we propose a model for the collective pili operation of N. gonorrheae bacteria that explains the experimentally observed behavior. Individual pili function independently but can lead to coordinated motion or pausing via the force balance. The geometry of the cell defines its orientation during motion. We show that by changing pili substrate interactions, the motility pattern can be altered in a predictable way. Although the model proposed is tangibly simple, it still has sufficient robustness to incorporate further advanced pili features and various cell geometries to describe other bacteria that employ pili to move on surfaces.     

Precise particle tracking against a complicated background: polynomial fitting with Gaussian weight

We present a new particle tracking software algorithm designed to accurately track the motion of low-contrast particles against a background with large variations in light levels. The method is based on a polynomial fit of the intensity around each feature point, weighted by a Gaussian function of the distance from the centre, and is especially suitable for tracking endogeneous particles in the cell, imaged with bright field, phase contrast or fluorescence optical microscopy. Furthermore, the method can simultaneously track particles of all different sizes, and allows significant freedom in their shape. The algorithm is evaluated using the quantitative measures of accuracy and precision of previous authors, using simulated images at variable signal-to-noise ratios. To these we add new tests: the error due to a non-uniform background, and the error due to two particles approaching each other. Finally the tracking of particles in real cell images is demonstrated. The method is made freely available for non-commercial use as a software package with a graphical user-interface, which can be run within the Matlab programming environment.     

The Development of Vision in the Zebrafish (Danio rerio)

We studied the development and maturation of the visual system by determining when zebrafish begin to see and to move their eyes. This information was correlated with the time courses of the development of the retina, the retinofugal projection, the retinal image, and the extraocular muscles, to obtain an integrated picture of early visual development. Two visual behaviors were monitored over 48–96 hr postfertilization (hpf). The startle response (body twitch) was evoked by an abrupt decrease in light intensity. The optokinetic response (tracking eye movements) was evoked by rotation of a striped drum. Visually evoked startle developed over 68–79 hpf, more than 20 hr after the onset of a touch-evoked startle. It was not seen in eyeless fish, excluding a role for nonretinal light senses. Tracking eye movements developed over 73–80 hpf. They were always in the direction of drum rotation, even when the fish had been light deprived from blastula stage, ruling out a “trial and error” period of learning to track the drum. The image formed by the ocular lens was examined in intact fish made transparent by suppressing the formation of melanin. The eye was initially far sighted and gradually improved, so that by 72 hpf the image plane coincided with the photoreceptor layer. The extraocular muscles assumed their adult configuration between 66 and 72 hpf. Thus, the retinal image and functional extraocular muscles appeared nearly simultaneously with the onset of tracking eye movements and probably represent the last events in the construction of this behavior.     

Automatic detection of fish and tracking of movement for ecology

1. Animal movement studies are conducted to monitor ecosystem health, understand ecological dynamics, and address management and conservation questions. In marine environments, traditional sampling and monitoring methods to measure animal movement are invasive, labor intensive, costly, and limited in the number of individuals that can be feasibly tracked. Automated detection and tracking of small‐scale movements of many animals through cameras are possible but are largely untested in field conditions, hampering applications to ecological questions.
2. Here, we aimed to test the ability of an automated object detection and object tracking pipeline to track small‐scale movement of many individuals in videos. We applied the pipeline to track fish movement in the field and characterize movement behavior. We automated the detection of a common fisheries species (yellowfin bream, Acanthopagrus australis) along a known movement passageway from underwater videos. We then tracked fish movement with three types of tracking algorithms (MOSSE, Seq‐NMS, and SiamMask) and evaluated their accuracy at characterizing movement.
3. We successfully detected yellowfin bream in a multispecies assemblage (F1 score =91%). At least 120 of the 169 individual bream present in videos were correctly identified and tracked. The accuracies among the three tracking architectures varied, with MOSSE and SiamMask achieving an accuracy of 78% and Seq‐NMS 84%.
4. By employing this integrated object detection and tracking pipeline, we demonstrated a noninvasive and reliable approach to studying fish behavior by tracking their movement under field conditions. These cost‐effective technologies provide a means for future studies to scale‐up the analysis of movement across many visual monitoring systems.

     

Filtering and interpreting location errors in satellite telemetry of marine animals

Argos telemetry offers a powerful means of tracking wild animals in their habitat, yet the delivered locations are subject to complex errors and random coverage. Bayesian filters and statistical models allow for objective trajectory estimates and inference on movement rates. As an alternative to Monte-Carlo methods, we investigate here how classic time series technique, such as the Kalman Filter, can be made robust to uncover patterns in the data. Our approach relies on a composite measurement model to account for outliers, and makes use of all the Location Classes to smooth observations and regularize the track to a regular time grid. Two application examples are presented. Using data from freely-swimming leatherback turtles, we confirm that locations of class A (LCA) are more accurate on average than class 0, and we recommend their use in tracking studies. We further show how measurement errors (and their geometry) interact with the assumed movement model, further modulating the final location error and the discriminating ability of the filter. The choice of the movement model appears important, since a model with no velocity constraint may fit observational errors at the expense of trajectory smoothness, while a speed-based model is better behaved but less forgiving for data fitting and outlier identification. Varying sea surface temperatures also appear to degrade the quality of locations and increase the occurrence of outliers, possibly in relation to thermal stratification and depth behavior. These results have important implications when inferring changes in behavior from long-term movements.     

Born Knowing: Tentacled Snakes Innately Predict Future Prey Behavior

Background

Aquatic tentacled snakes (Erpeton tentaculatus) can take advantage of their prey''s escape response by startling fish with their body before striking. The feint usually startles fish toward the snake''s approaching jaws. But when fish are oriented at a right angle to the jaws, the C-start escape response translates fish parallel to the snake''s head. To exploit this latter response, snakes must predict the future location of the fish. Adult snakes can make this prediction. Is it learned, or are tentacled snakes born able to predict future fish behavior?

Methods and Findings

Laboratory-born, naïve snakes were investigated as they struck at fish. Trials were recorded at 250 or 500 frames per second. To prevent learning, snakes were placed in a water container with a clear transparency sheet or glass bottom. The chamber was placed over a channel in a separate aquarium with fish below. Thus snakes could see and strike at fish, without contact. The snake''s body feint elicited C-starts in the fish below the transparency sheet, allowing strike accuracy to be quantified in relationship to the C-starts. When fish were oriented at a right angle to the jaws, naïve snakes biased their strikes to the future location of the escaping fish''s head, such that the snake''s jaws and the fish''s translating head usually converged. Several different types of predictive strikes were observed.

Conclusions

The results show that some predators have adapted their nervous systems to directly compensate for the future behavior of prey in a sensory realm that usually requires learning. Instead of behavior selected during their lifetime, newborn tentacled snakes exhibit behavior that has been selected on a different scale—over many generations. Counter adaptations in fish are not expected, as tentacled snakes are rare predators exploiting fish responses that are usually adaptive.     

Do juvenile gape-limited predators compensate for their small size when feeding?

When juvenile and adult animals occur syntopically, juveniles are at a distinct performance disadvantage due to their absolutely small size. Yet, optimal foraging theory predicts that juvenile predators should feed efficiently in order to compete with adults for food, and to minimize their exposure to predators. Previous authors have suggested that one way for juvenile animals to accomplish these ecological tasks is by increasing their overall feeding performance relative to adults (compensation hypothesis). Nonetheless, only a handful of studies have tested whether juvenile animals have increased feeding performance (e.g. decreased ingestion and/or handling times relative to body size) compared with adults. We tested this hypothesis by examining the ontogeny of head dimensions and feeding performance (ingestion time and number of mandibular protractions) on fish prey for broad-banded water snakes Nerodia fasciata . Individuals were fed fish scaled in a 1:1 ratio to their head width. All head dimensions scaled with significant negative allometry versus body size, and thus smaller snakes had relatively larger heads for their body size compared with larger snakes. By contrast, most head variables (except head volume) exhibited positive allometry versus head length, demonstrating that larger snakes had larger head dimensions relative to head size compared with smaller snakes. In the performance trials, smaller snakes had worse feeding performances when feeding on similarly sized fish prey (relative to their head width) compared with larger snakes. Therefore, these data show that smaller water snakes do not compensate for their size through increased feeding performance.     

眼斑双锯鱼仔稚鱼发育异速生长

运用生态学和传统理论生物学的研究方法, 对孵化后眼斑双锯(Amphiprion ocellaris)仔、稚鱼在早期生存和环境适应上的异速生长及器官优先发育生态学意义进行了研究, 以期为眼斑双锯鱼人工繁殖和育苗提供参考资料。以11日龄为眼斑双锯鱼仔、稚鱼的区分时期, 结果表明, 眼斑双锯鱼仔、稚鱼的感觉、摄食和游泳等器官快速分化, 均存在异速生长现象。在头部器官中, 吻长、眼间距、口宽和头高在仔鱼期均为正异速生长, 吻至鳃裂前缘长和眼径为负异速生长。在身体各部位中, 仔鱼期体高、躯干长、尾长、尾柄长、尾柄高和体厚均为正异速生长, 仅头长为负异速生长; 在游泳器官中, 仔鱼期眼斑双锯鱼尾鳍、背鳍、胸鳍、腹鳍和臀鳍均为正异速生长。稚鱼期眼斑双锯鱼头部、躯干及游泳等各器官均为负异速生长。眼斑双锯鱼这些关键器官的异速发育, 对适应环境因子变化具有重要的生态学意义。     

Parasitic freshwater pearl mussel larvae (Margaritifera margaritifera L.) reduce the drift-feeding rate of juvenile brown trout (Salmo trutta L.)

In this paper we describe, for the first time, the effects of freshwater pearl mussel (Margaritifera margaritifera L.) encystment on the drift-feeding behavior of juvenile brown trout (Salmo trutta L.). Because both mussel and salmonid populations are often threatened, this study not only adds knowledge to the understanding of host-parasite systems, but it is also of conservation value. Individual trout, mussel-encysted (25.1?±?5.7 larvae?·?g^?1 body weight, n?=?5) or non-encysted (n?=?5), were fed with chironomid larvae in a flow-through stream aquarium. Feeding trials were filmed and analyzed by counting the numbers of chironomid larvae each individual ate, and by estimating the prey-capture distance. Non-encysted trout had a significantly higher drift-foraging rate than did encysted trout, and they captured significantly more prey further away from their focal point. The reduced foraging success of encysted trout was mainly due to their failure to catch prey relatively further from their focal point. This suggests that reduced foraging success of encysted trout may be due to poorer energetic status, but the physical effects of mussel larvae on prey handling time cannot be ruled out. Encysted trout caught approximately 20 % fewer prey, which would result in a reduction in growth potential during the period of mussel encystment. Reduced energetic status might also result in reduced competitive ability or in increased exposure to predation risk.     

The dynamics of coordinated group hunting and collective information transfer among schooling prey

Predator-prey interactions are vital to the stability of many ecosystems. Yet, few studies have considered how they are mediated due to substantial challenges in quantifying behavior over appropriate temporal and spatial scales. Here, we employ high-resolution sonar imaging to track the motion and interactions among predatory fish and their schooling prey in a natural environment. In particular, we address the relationship between predator attack behavior and the capacity for prey to respond both directly and through collective propagation of changes in velocity by group members. To do so, we investigated a large number of attacks and estimated per capita risk during attack and its relation to the size, shape, and internal structure of prey groups. Predators were found to frequently form coordinated hunting groups, with up to five individuals attacking in line formation. Attacks were associated with increased fragmentation and irregularities in the spatial structure of prey groups, features that inhibit collective information transfer among prey. Prey group fragmentation, likely facilitated by predator line formation, increased (estimated) per capita risk of prey, provided prey schools were maintained below a threshold size of approximately 2 m(2). Our results highlight the importance of collective behavior to the strategies employed by both predators and prey under conditions of considerable informational constraints.     

The effect of body coloration and group size on social partner preferences in female fighting fish (Betta splendens)

Females of the fighting fish Betta splendens have been shown to associate with other B. splendens females in a manner reminiscent of shoaling behavior. Since body coloration varies dramatically in this species, and since body coloration has been shown to affect shoalmate choice in other species of fish, we examined the influence of body coloration on association preferences in female B. splendens. In dichotomous choice tests, B. splendens females spent more time swimming near groups of females (regardless of coloration) than swimming near an empty chamber, and chose to swim near fish of similar coloration to their own when choosing between two distinctly colored groups of females. When examining the interplay between body coloration and group size, focal fish spent more time swimming near larger groups (N = 5) of similarly colored fish than swimming near an individual female of similar coloration. However, focal fish showed no preference when presented with an individual female of similar coloration and a larger group of females of dissimilar coloration. These results suggest that association choices in B. splendens females are strongly affected by both body coloration and by group size.     

Visual fields, eye movements, and scanning behavior of a sit-and-wait predator, the black phoebe (Sayornis nigricans)

Foraging mode influences the dominant sensory modality used by a forager and likely the strategies of information gathering used in foraging and anti-predator contexts. We assessed three components of visual information gathering in a sit-and-wait avian predator, the black phoebe (Sayornis nigricans): configuration of the visual field, degree of eye movement, and scanning behavior through head-movement rates. We found that black phoebes have larger lateral visual fields than similarly sized ground-foraging passerines, as well as relatively narrower binocular and blind areas. Black phoebes moved their eyes, but eye movement amplitude was relatively smaller than in other passerines. Black phoebes may compensate for eye movement constraints with head movements. The rate of head movements increased before attacking prey in comparison to non-foraging contexts and before movements between perches. These findings suggest that black phoebes use their lateral visual fields, likely subtended by areas of high acuity in the retina, to track prey items in a three-dimensional space through active head movements. These head movements may increase depth perception, motion detection and tracking. Studying information gathering through head movement changes, rather than body posture changes (head-up, head-down) as generally presented in the literature, may allow us to better understand the mechanisms of information gathering from a comparative perspective.     

Collective dynamics of elastically coupled Myosin v motors

Characterization of the collective behaviors of different classes of processive motor proteins has become increasingly important to understand various intracellular trafficking and transport processes. This work examines the dynamics of structurally-defined motor complexes containing two myosin Va (myoVa) motors that are linked together via a molecular scaffold formed from a single duplex of DNA. Dynamic changes in the filament-bound configuration of these complexes due to motor binding, stepping, and detachment were monitored by tracking the positions of different color quantum dots that report the position of one head of each myoVa motor on actin. As in studies of multiple kinesins, the run lengths produced by two myosins are only slightly larger than those of single motor molecules. This suggests that internal strain within the complexes, due to asynchronous motor stepping and the resultant stretching of motor linkages, yields net negative cooperative behaviors. In contrast to multiple kinesins, multiple myosin complexes move with appreciably lower velocities than a single-myosin molecule. Although similar trends are predicted by a discrete state stochastic model of collective motor dynamics, these analyses also suggest that multiple myosin velocities and run lengths depend on both the compliance and the effective size of their cargo. Moreover, it is proposed that this unique collective behavior occurs because the large step size and relatively small stalling force of myoVa leads to a high sensitivity of motor stepping rates to strain.