| 1. |
|
|
| 2. |
- Loni, Mohammad, et al.
(författare)
-
Designing compact convolutional neural network for embedded stereo vision systems
- 2018
-
Ingår i: Proceedings - 2018 IEEE 12th International Symposium on Embedded Multicore/Many-Core Systems-on-Chip, MCSoC 2018. - : Institute of Electrical and Electronics Engineers (IEEE). - 9781538666890 ; , s. 244-251
-
Konferensbidrag (refereegranskat)abstract
- Autonomous systems are used in a wide range of domains from indoor utensils to autonomous robot surgeries and self-driving cars. Stereo vision cameras probably are the most flexible sensing way in these systems since they can extract depth, luminance, color, and shape information. However, stereo vision based applications suffer from huge image sizes and computational complexity leading system to higher power consumption. To tackle these challenges, in the first step, GIMME2 stereo vision system [1] is employed. GIMME2 is a high-throughput and cost efficient FPGA-based stereo vision embedded system. In the next step, we present a framework for designing an optimized Deep Convolutional Neural Network (DCNN) for time constraint applications and/or limited resource budget platforms. Our framework tries to automatically generate a highly robust DCNN architecture for image data receiving from stereo vision cameras. Our proposed framework takes advantage of a multi-objective evolutionary optimization approach to design a near-optimal network architecture for both the accuracy and network size objectives. Unlike recent works aiming to generate a highly accurate network, we also considered the network size parameters to build a highly compact architecture. After designing a robust network, our proposed framework maps generated network on a multi/many core heterogeneous System-on-Chip (SoC). In addition, we have integrated our framework to the GIMME2 processing pipeline such that it can also estimate the distance of detected objects. The generated network by our framework offers up to 24x compression rate while losing only 5% accuracy compare to the best result on the CIFAR-10 dataset.
|
|
| 3. |
- Karlsson, E., et al.
(författare)
-
Scaffolding of a bacterial genome using MinION nanopore sequencing
- 2015
-
Ingår i: Scientific Reports. - : Springer Science and Business Media LLC. - 2045-2322. ; 5
-
Tidskriftsartikel (refereegranskat)abstract
- Second generation sequencing has revolutionized genomic studies. However, most genomes contain repeated DNA elements that are longer than the read lengths achievable with typical sequencers, so the genomic order of several generated contigs cannot be easily resolved. A new generation of sequencers offering substantially longer reads is emerging, notably the Pacific Biosciences (PacBio) RS II system and the MinION system, released in early 2014 by Oxford Nanopore Technologies through an early access program. The latter has highly advantageous portability and sequences samples by measuring changes in ionic current when single-stranded DNA molecules are translocated through nanopores. We show that the MinION system produces long reads with high mapability that can be used for scaffolding bacterial genomes, despite currently producing substantially higher error rates than PacBio reads. With further development we anticipate that MinION will be useful not only for assembling genomes, but also for rapid detection of organisms, potentially in the field.
|
|
| 4. |
- Melin, Anna K., Assistant Professor, 1965-, et al.
(författare)
-
Energy availability and the female athlete triad in elite endurance athletes
- 2015
-
Ingår i: Scandinavian Journal of Medicine and Science in Sports. - : Wiley-Blackwell. - 0905-7188 .- 1600-0838. ; 25:5, s. 610-622
-
Tidskriftsartikel (refereegranskat)abstract
- The female athlete triad (Triad), links low energy availability (EA), with menstrual dysfunction (MD), and impaired bone health. The aims of this study were to examine associations between EA/MD and energy metabolism and the prevalence of Triad-associated conditions in endurance athletes. Forty women [26.2 +/- 5.5 years, body mass index (BMI) 20.6 +/- 2.0 kg/m(2), body fat 20.0 +/- 3.0%], exercising 11.4 +/- 4.5 h/week, were recruited from national teams and competitive clubs. Protocol included gynecological examination; assessment of bone health; indirect respiratory calorimetry; diet and exercise measured 7 days to assess EA; eating disorder (ED) examination; blood analysis. Subjects with low/reduced EA (< 45 kcal/kg FFM/day), had lower resting metabolic rate (RMR) compared with those with optimal EA [28.4 +/- 2.0 kcal/kg fat-free mass (FFM)/day vs 30.5 +/- 2.2 kcal/kg FFM/day, P < 0.01], as did subjects with MD compared with eumenorrheic subjects (28.6 +/- 2.4 kcal/kg FFM/day vs 30.2 +/- 1.8 kcal/kg FFM/day, P < 0.05). 63% had low/reduced EA, 25% ED, 60% MD, 45% impaired bone health, and 23% had all three Triad conditions. 53% had low RMR, 25% hypercholesterolemia, and 38% hypoglycemia. Conclusively, athletes with low/reduced EA and/or MD had lowered RMR. Triad-associated conditions were common in this group of athletes, despite a normal BMI range. The high prevalence of ED, MD, and impaired bone health emphasizes the importance of prevention, early detection, and treatment of energy deficiency.
|
|
| 5. |
- Owers, Katharine A., et al.
(författare)
-
Adaptation to infectious disease exposure in indigenous Southern African populations
- 2017
-
Ingår i: Proceedings of the Royal Society of London. Biological Sciences. - : ROYAL SOC. - 0962-8452 .- 1471-2954. ; 284:1852
-
Tidskriftsartikel (refereegranskat)abstract
- Genetic analyses can provide information about human evolutionary history that cannot always be gleaned from other sources. We evaluated evidence of selective pressure due to introduced infectious diseases in the genomes of two indigenous southern African San groups-the double dagger Khomani who had abundant contact with other people migrating into the region and the more isolated Ju vertical bar'hoansi. We used a dual approach to test for increased selection on immune genes compared with the rest of the genome in these groups. First, we calculated summary values of statistics that measure genomic signatures of adaptation to contrast selection signatures in immune genes and all genes. Second, we located regions of the genome with extreme values of three selection statistics and examined these regions for enrichment of immune genes. We found stronger and more abundant signals of selection in immune genes in the double dagger Khomani than in the Ju vertical bar'hoansi. We confirm this finding within each population to avoid effects of different demographic histories of the two populations. We identified eight immune genes that have potentially been targets of strong selection in the double dagger Khomani, whereas in the Juj'hoansi, no immune genes were found in the genomic regions with the strongest signals of selection. We suggest that the more abundant signatures of selection at immune genes in the double dagger Khomani could be explained by their more frequent contact with immigrant groups, which likely led to increased exposure and adaptation to introduced infectious diseases.
|
|
