| 1. |
- Amundin, Mats, et al.
(författare)
-
A proposal to use distributional models to analyse dolphin vocalisation
- 2017
-
Ingår i: Proceedings of the 1st International Workshop on Vocal Interactivity in-and-between Humans, Animals and Robots, VIHAR 2017. - 9782956202905 ; , s. 31-32
-
Konferensbidrag (refereegranskat)abstract
- This paper gives a brief introduction to the starting points of an experimental project to study dolphin communicative behaviour using distributional semantics, with methods implemented for the large scale study of human language.
|
|
| 2. |
- Haubro Andersen, Pia, et al.
(författare)
-
Towards Machine Recognition of Facial Expressions of Pain in Horses
- 2021
-
Ingår i: Animals. - : MDPI. - 2076-2615. ; 11:6
-
Forskningsöversikt (refereegranskat)abstract
- Simple Summary Facial activity can convey valid information about the experience of pain in a horse. However, scoring of pain in horses based on facial activity is still in its infancy and accurate scoring can only be performed by trained assessors. Pain in humans can now be recognized reliably from video footage of faces, using computer vision and machine learning. We examine the hurdles in applying these technologies to horses and suggest two general approaches to automatic horse pain recognition. The first approach involves automatically detecting objectively defined facial expression aspects that do not involve any human judgment of what the expression "means". Automated classification of pain expressions can then be done according to a rule-based system since the facial expression aspects are defined with this information in mind. The other involves training very flexible machine learning methods with raw videos of horses with known true pain status. The upside of this approach is that the system has access to all the information in the video without engineered intermediate methods that have filtered out most of the variation. However, a large challenge is that large datasets with reliable pain annotation are required. We have obtained promising results from both approaches. Automated recognition of human facial expressions of pain and emotions is to a certain degree a solved problem, using approaches based on computer vision and machine learning. However, the application of such methods to horses has proven difficult. Major barriers are the lack of sufficiently large, annotated databases for horses and difficulties in obtaining correct classifications of pain because horses are non-verbal. This review describes our work to overcome these barriers, using two different approaches. One involves the use of a manual, but relatively objective, classification system for facial activity (Facial Action Coding System), where data are analyzed for pain expressions after coding using machine learning principles. We have devised tools that can aid manual labeling by identifying the faces and facial keypoints of horses. This approach provides promising results in the automated recognition of facial action units from images. The second approach, recurrent neural network end-to-end learning, requires less extraction of features and representations from the video but instead depends on large volumes of video data with ground truth. Our preliminary results suggest clearly that dynamics are important for pain recognition and show that combinations of recurrent neural networks can classify experimental pain in a small number of horses better than human raters.
|
|
| 3. |
- Svartström, Olov, et al.
(författare)
-
Ninety-nine de novo assembled genomes from the moose (Alces alces) rumen microbiome provide new insights into microbial plant biomass degradation
- 2017
-
Ingår i: The ISME Journal. - : Nature Publishing Group. - 1751-7362 .- 1751-7370. ; 11:11, s. 2538-2551
-
Tidskriftsartikel (refereegranskat)abstract
- The moose (Alces alces) is a ruminant that harvests energy from fiber-rich lignocellulose material through carbohydrate-active enzymes (CAZymes) produced by its rumen microbes. We applied shotgun metagenomics to rumen contents from six moose to obtain insights into this microbiome. Following binning, 99 metagenome-assembled genomes (MAGs) belonging to 11 prokaryotic phyla were reconstructed and characterized based on phylogeny and CAZyme profile. The taxonomy of these MAGs reflected the overall composition of the metagenome, with dominance of the phyla Bacteroidetes and Firmicutes. Unlike in other ruminants, Spirochaetes constituted a significant proportion of the community and our analyses indicate that the corresponding strains are primarily pectin digesters. Pectin-degrading genes were also common in MAGs of Ruminococcus, Fibrobacteres and Bacteroidetes and were overall overrepresented in the moose microbiome compared with other ruminants. Phylogenomic analyses revealed several clades within the Bacteriodetes without previously characterized genomes. Several of these MAGs encoded a large numbers of dockerins, a module usually associated with cellulosomes. The Bacteroidetes dockerins were often linked to CAZymes and sometimes encoded inside polysaccharide utilization loci, which has never been reported before. The almost 100 CAZyme-annotated genomes reconstructed in this study provide an in-depth view of an efficient lignocellulose-degrading microbiome and prospects for developing enzyme technology for biorefineries.
|
|
| 4. |
- Klütsch, Cornelya, et al.
(författare)
-
Regional occurrence, high frequency but low diversity of mitochondrial DNA haplogroup d1 suggests a recent dog-wolf hybridization in Scandinavia
- 2011
-
Ingår i: Animal Genetics. - : Wiley. - 0268-9146 .- 1365-2052. ; 42:1, s. 100-103
-
Tidskriftsartikel (refereegranskat)abstract
- P>The domestic dog mitochondrial DNA (mtDNA)-gene pool consists of a homogenous mix of haplogroups shared among all populations worldwide, indicating that the dog originated at a single time and place. However, one small haplogroup, subclade d1, found among North Scandinavian/Finnish spitz breeds at frequencies above 30%, has a clearly separate origin. We studied the genetic and geographical diversity for this phylogenetic group to investigate where and when it originated and whether through independent domestication of wolf or dog-wolf crossbreeding. We analysed 582 bp of the mtDNA control region for 514 dogs of breeds earlier shown to harbour d1 and possibly related northern spitz breeds. Subclade d1 occurred almost exclusively among Swedish/Finnish Sami reindeer-herding spitzes and some Swedish/Norwegian hunting spitzes, at a frequency of mostly 60-100%. Genetic diversity was low, with only four haplotypes: a central, most frequent, one surrounded by two haplotypes differing by an indel and one differing by a substitution. The substitution was found in a single lineage, as a heteroplasmic mix with the central haplotype. The data indicate that subclade d1 originated in northern Scandinavia, at most 480-3000 years ago and through dog-wolf crossbreeding rather than a separate domestication event. The high frequency of d1 suggests that the dog-wolf hybrid phenotype had a selective advantage.
|
|
| 5. |
- Broomé, Sofia, et al.
(författare)
-
Dynamics are important for the recognition of equine pain in video
- 2019
-
Ingår i: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. - : Institute of Electrical and Electronics Engineers (IEEE). - 1063-6919.
-
Konferensbidrag (refereegranskat)abstract
- A prerequisite to successfully alleviate pain in animals is to recognize it, which is a great challenge in non-verbal species. Furthermore, prey animals such as horses tend to hide their pain. In this study, we propose a deep recurrent two-stream architecture for the task of distinguishing pain from non-pain in videos of horses. Different models are evaluated on a unique dataset showing horses under controlled trials with moderate pain induction, which has been presented in earlier work. Sequential models are experimentally compared to single-frame models, showing the importance of the temporal dimension of the data, and are benchmarked against a veterinary expert classification of the data. We additionally perform baseline comparisons with generalized versions of state-of-the-art human pain recognition methods. While equine pain detection in machine learning is a novel field, our results surpass veterinary expert performance and outperform pain detection results reported for other larger non-human species.
|
|
| 6. |
- Pohl, Johannes, et al.
(författare)
-
Effects of ozonated sewage effluent on reproduction and behavioral endpoints in zebrafish (Danio rerio)
- 2018
-
Ingår i: Aquatic Toxicology. - : Elsevier BV. - 0166-445X .- 1879-1514. ; 200, s. 93-101
-
Tidskriftsartikel (refereegranskat)abstract
- Pharmaceutical residues and other micro-contaminants may enter aquatic environments through effluent from sewage treatment plants (STPs) and could cause adverse effects in wild fish. One strategy to alleviate this situation is to improve wastewater treatment by ozonation. To test the effectiveness of full-scale wastewater effluent ozonation at a Swedish municipal STP, the added removal efficiency was measured for 105 pharmaceuticals. In addition, gene expression, reproductive and behavioral endpoints were analyzed in zebrafish (Danio rerio) exposed on-site over 21 days to ozonated or non-ozonated effluents as well as to tap water. Ozone treatment (7 g O-3/m(3)) removed pharmaceuticals by an average efficiency of 77% in addition to the conventional treatment, leaving 11 screened pharmaceuticals above detection limits. Differences in biological responses of the exposure treatments were recorded in gene expression, reproduction and behavior. Hepatic vitellogenin gene expression was higher in male zebrafish exposed to the ozonated effluent compared to the non-ozonated effluent and tap water treatments. The reproductive success was higher in fish exposed to ozonated effluent compared to non-ozonated effluent and to tap water. The behavioral measurements showed that fish exposed to the ozonated STP effluent were less active in swimming the first minute after placed in a novel vessel. Ozonation is a capable method for removing pharmaceuticals in effluents. However, its implementation should be thoroughly evaluated for any potential biological impact. Future research is needed for uncovering the factors which produced the in vivo responses in fish.
|
|
| 7. |
- Rashid, Maheen, et al.
(författare)
-
Equine Pain Behavior Classification via Self-Supervised Disentangled Pose Representation
- 2022
-
Ingår i: 2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022). - : Institute of Electrical and Electronics Engineers (IEEE). ; , s. 152-162
-
Konferensbidrag (refereegranskat)abstract
- Timely detection of horse pain is important for equine welfare. Horses express pain through their facial and body behavior, but may hide signs of pain from unfamiliar human observers. In addition, collecting visual data with detailed annotation of horse behavior and pain state is both cumbersome and not scalable. Consequently, a pragmatic equine pain classification system would use video of the un-observed horse and weak labels. This paper proposes such a method for equine pain classification by using multi-view surveillance video footage of unobserved horses with induced orthopaedic pain, with temporally sparse video level pain labels. To ensure that pain is learned from horse body language alone, we first train a self-supervised generative model to disentangle horse pose from its appearance and background before using the disentangled horse pose latent representation for pain classification. To make best use of the pain labels, we develop a novel loss that formulates pain classification as a multi-instance learning problem. Our method achieves pain classification accuracy better than human expert performance with 60% accuracy. The learned latent horse pose representation is shown to be viewpoint covariant, and disentangled from horse appearance. Qualitative analysis of pain classified segments shows correspondence between the pain symptoms identified by our model, and equine pain scales used in veterinary practice.
|
|
| 8. |
- Falkenberg Hansen, Kjetil, 1972-
(författare)
-
The acoustics and performance of DJ scratching, Analysis and modelling
- 2010
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis focuses on the analysis and modeling of scratching, in other words, the DJ (disk jockey) practice of using the turntable as a musical instrument. There has been experimental use of turntables as musical instruments since their invention, but the use is now mainly ascribed to the musical genre hip-hop and the playing style known as scratching. Scratching has developed to become a skillful instrument-playing practice with complex musical output performed by DJs. The impact on popular music culture has been significant, and for many, the DJ set-up of turntables and a mixer is now a natural instrument choice for undertaking a creative music activity. Six papers are included in this thesis, where the first three approach the acoustics and performance of scratching, and the second three approach scratch modeling and the DJ interface. Additional studies included here expand on the scope of the papers. For the acoustics and performance studies, DJs were recorded playing both demonstrations of standard performance techniques, and expressive performances on sensor-equipped instruments. Analysis of the data revealed that there are both differences and commonalities in playing strategies between musicians, and between expressive intentions. One characteristic feature of scratching is the range of standard playing techniques, but in performances it seems DJs vary the combination of playing techniques more than the rendering of these techniques. The third study describes some of the acoustic parameters of typical scratch improvisations and looks at which musical parameters are typically used for expressive performances. Extracted acoustic and performance parameters from the data show the functional ranges within which DJs normally play. Unlike traditional musical instruments, the equipment used for scratching was not intended to be used for creating music. The interface studies focus on traditional as well as new interfaces for DJs, where parameter mappings between input gestures and output signal are described. Standard performance techniques have been modeled in software called Skipproof, based on results from the first papers. Skipproof was used for testing other types of controllers than turntables, where complex DJ gestures could be manipulated using simplified control actions, enabling even non-experts to play expressively within the stylistic boundaries of DJ scratching. The last paper describes an experiment of using an existing hardware platform, the Reactable, to help designing and prototyping the interaction between different sound models and instrument interfaces, including scratching and Skipproof. In addition to the included papers, studies were conducted of expressivity, description of the emotional contents of scratching, DJ playing activities, and the coupling between playing techniques and sample. The physical affordances of the turntable, mixer and samples, as well as genre conventions of hip-hop, are assumed to explain some of the findings that distinguish scratching from other instrumental sounds or practices.
|
|
| 9. |
- Pang, Jun-Feng, et al.
(författare)
-
mtDNA data indicate a single origin for dogs south of Yangtze River, less than 16,300 years ago, from numerous wolves.
- 2009
-
Ingår i: Molecular biology and evolution. - : Oxford University Press (OUP). - 1537-1719 .- 0737-4038. ; 26:12, s. 2849-64
-
Tidskriftsartikel (refereegranskat)abstract
- There is no generally accepted picture of where, when, and how the domestic dog originated. Previous studies of mitochondrial DNA (mtDNA) have failed to establish the time and precise place of origin because of lack of phylogenetic resolution in the so far studied control region (CR), and inadequate sampling. We therefore analyzed entire mitochondrial genomes for 169 dogs to obtain maximal phylogenetic resolution and the CR for 1,543 dogs across the Old World for a comprehensive picture of geographical diversity. Hereby, a detailed picture of the origins of the dog can for the first time be suggested. We obtained evidence that the dog has a single origin in time and space and an estimation of the time of origin, number of founders, and approximate region, which also gives potential clues about the human culture involved. The analyses showed that dogs universally share a common homogenous gene pool containing 10 major haplogroups. However, the full range of genetic diversity, all 10 haplogroups, was found only in southeastern Asia south of Yangtze River, and diversity decreased following a gradient across Eurasia, through seven haplogroups in Central China and five in North China and Southwest (SW)Asia, down to only four haplogroups in Europe. The mean sequence distance to ancestral haplotypes indicates an origin 5,400-16,300 years ago (ya) from at least 51 female wolf founders. These results indicate that the domestic dog originated in southern China less than 16,300 ya, from several hundred wolves. The place and time coincide approximately with the origin of rice agriculture, suggesting that the dogs may have originated among sedentary hunter-gatherers or early farmers, and the numerous founders indicate that wolf taming was an important culture trait.
|
|
| 10. |
|
|
