Whether it's capturing a car's license plate on the highway or detecting someone's facial features to tag friends, computer vision and image processing have found their way into many facets of our lives. Image and video processing algorithms ultimately tailor towards one of two goals: to analyze data and produce output in as close to real-time as possible, or to take in and operate on large swaths of information offline. Image convolution is a mathematical method with which we can filter an image to highlight or make clearer desired information. The most popular uses of image convolution accentuate edges, corners, and facial features for analysis. The goal of this project was to investigate various image convolution algorithms and compare them in terms of hardware usage, power utilization, and ability to handle substantial amounts of data in a reasonable amount of time. The algorithms were designed, simulated, and synthesized for the Zynq-7000 FPGA, selected both for its flexibility and low power consumption.

No description available.

Globalization of semiconductor design and manufacturing has led to a concern of trust in the final product. The components may now be designed and manufactured from anywhere in the world without the direct supervision of the buyer. As a result, the hardware designs and fabricated chips may be vulnerable to malicious alterations by an adversary at any stage of VLSI design flow, thus compromising the integrity of the component. The effect of any modifications made by the adversary can be catastrophic in the critical applications. Because of the stealthy nature of such insertions, it is extremely difficult to detect them using traditional testing and verification methods. Therefore, the trust of the hardware systems require a new approach and have drawn much attention in the hardware security community. For many years, the researchers have developed sophisticated techniques to detect, isolate and prevent malicious attacks in cyber security community assuming that the underlying hardware platform is extremely secure and trustworthy. But the hardware may contain one or more backdoors that can be exploited by software at the time of operation. Therefore, the trust of the computing system cannot be guaranteed unless we can guarantee the trust of the hardware platform. A malicious insertion can be very stealthy and may only involve minor modification in the hardware design or the fabricated chip. The insertion may require rare or specific conditions in order to be activated. The effect may be denial of service, change of function, destruction of chip, leakage of secret information from cryptographic hardware etc. In this thesis, we propose a novel technique for the detection of malicious alteration(s) in a third party soft intellectual property (IP) using a clever combination of sequential equivalence checking (SEC) and automatic test generation. The use of powerful inductive invariants can prune a large illegal state space, and test generation helps to provide a sensitization path for nodes of interest. Results for a set of hard-to-verify designs show that our method can either ensure that the suspect design is free from the functional effect of any malicious change(s) or return a small group of most likely malicious signals. / Master of Science

In this thesis, we present a new register-transfer level (RTL) test generation method that makes use of two coverage metrics, Branch Coverage, and Mutation Coverage across two stages, to cover hard-to-reach points previously unreached. We start with a preprocessing stage by converting the RTL source to a C++ equivalent using a modified Verilator, which also automatically creates mutants and the corresponding mutated C++ design, based on arithmetic, logical and relational operators during conversion. With the help of extracted Data Dependency and Control Flow Graphs, in every <golden, mutation> pair, branches containing variables dependent on the mutated statement are instrumented to track them. The first stage uses Evolutionary algorithms with Ant Colony Optimization to generate test vectors with mutation coverage as the metric. Two new filtering techniques are also proposed which optimize the first stage by eliminating the need for generating tests for redundant mutants. The next stage is the original BEACON which now takes the generated mutation test vectors as the initial population unlike random vectors, and output final test vectors. These test vectors succeed in improving the coverage up to 70%, compared to the previous approaches for most of the ITC99 benchmarks. With the application of filtering techniques, we also observed a speedup by 85% in the test generation runtime and also up to 78% reduction in test vector size when compared with those generated by the previous techniques. / MS / In the recent years, Verification has become one of the major bottlenecks in integrated circuit design process, which is exacerbated by the increasing design complexities today. Today designers start the design process by abstracting the initial design in a manner similar to software programming language using a higher abstraction language called Hardware Descriptive Language(HDL). Hence, an HDL based design also contains a number of case statements and if-else statements, also called branches, similar to a software design. Branches indicate decision points in the design and high branch coverage based tests can give us an assurance that the design is properly exercised as compared to those given by randomly generated tests. In this thesis, we introduce a new test generation methodology which generates tests using the help of user introduced mutants to ensure higher branch coverage. Mutation testing is similar to a fault testing method, in which an error or a fault is deliberately introduced into the design and we check if the tests generated are able to detect the fault. An important property of a mutant is that: when a mutant is applied and if the mutated part of the design is exercised by the given test suite, then the following data and control flow path taken can be different from that taken on the original design. This important property along with proper guidance is used in our work to reach some branches which are difficult to cover by random test vectors, and this is the main basis of this thesis. Applying this method, we observed that the branch coverage increased with a decrease in test generation runtime and test vector length when compared to previously proposed techniques.

Early closure to functional correctness of the final chip has become a crucial success factor in the semiconductor industry. In this context, the tedious task of functional debugging poses a significant bottleneck in modern electronic design processes, where new problems related to debugging are constantly introduced and predominantly performed manually. This dissertation proposes methodologies that address two emerging debugging problems in modern design flows. First, it proposes a novel and automated triage framework for Register-Transfer-Level (RTL) debugging. The proposed framework employs clustering techniques to automate the grouping of a plethora of failures that occur during regression verification. Experiments demonstrate accuracy improvements of up to 40% compared to existing triage methodologies. Next, it introduces new techniques for Field Programmable Gate Array (FPGA) debugging that leverage reconfigurability to allow debugging to operate without iterative executions of computationally-intensive design re-synthesis tools. Experiments demonstrate productivity improvements of up to 30 x vs. conventional approaches.

A rainbow trout liver cell line, RTL-W1, was used to evaluate the toxic potential of a heavy fuel oil (HFO) HFO 7102, and its fractions, which together with the HFO are referred to as the oil samples. The fractions were F2, F3, F3-1, F3-2 and F4 and had been prepared by low-temperature vacuum distillation by collaborators at Queen's University. For presentation to the cells, HFO 7102 and its fractions were made into High Energy-Chemically Enhanced Water Accommodated Fractions (HE-CEWAFs). The procedure for this involved adding Corexit 9500 to the oil samples, mixing them on a vortex, and letting the phases settle. The HE-CEWAFs were added to RTL-W1 cell cultures, and at various times afterwards cell viability and CYP1A induction were monitored. Cell viability was evaluated with two dyes, Alamar Blue, which monitors energy metabolism, and 5-carboxfluorescein diacetate acetoxymethyl ester (CFDA AM), which measures plasma membrane integrity. With both indicator dyes, Corexit 9500 was cytotoxic but the concentrations eliciting cytotoxicity varied with the cell culture media. In Leibovitz's L-15 with fetal bovine serum (FBS), which was the medium used for studying CYP1A induction, Corexit 9500 was only cytotoxic at concentrations of 0.1 % (v/v) and greater. For the oil samples, F3-2 at 1 mg/ml and F4 at 10 mg/ml, which were the highest testable concentrations for each, no loss of cell viability was observed over 24 h. The other oil samples were cytotoxic only at their highest testable concentrations, which ended being between 1 and 10 mg/ml. CYP1A induction was monitored in RTL-W1 as catalytic activity and as the level of CYP1A (P4501A) protein. The catalytic activity was assayed as 7-ethoxyresorufin o-deethylase (EROD) activity; the CYP1A protein level, by western blotting. The positive control was 2, 3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), which strongly induced both EROD activity and CYP1A protein. Corexit 9500 by itself neither induced EROD activity nor CYP1A protein. All the oil samples induced both EROD activity and CYP1A protein. With both endpoints, the most potent fraction was F3; the least potent, F3-2. As the induction of CYP1A is associated with the development of blue sac disease (BSD) and mortality in early life stages of fish, the results suggest that HFO 7102 and its fractions have the potential to reduce recruitment of young into adult fish populations. CYP1A induction by F3 was studied further, again through EROD activity and western blotting. As the F3 concentration was increased, EROD activity increased but declined at high concentrations, whereas CYP1A protein continued to increase. This suggests the presence of compounds in F3 that at high concentrations inhibit the catalytic activity of CYP1A. When F3 was presented to RTL-W1 cultures together with TCDD, CYP1A protein was induced but not EROD activity. Again this suggests that F3 contains inhibitor(s) of CYP1A as well as inducers. When cultures were exposed to either F3 or TCDD for 24 h and then followed by western blotting for up to 6 days after F3 or TCDD removal, CYP1A levels declined in F3 cultures but not in TCDD cultures. This suggests that RTL-W1 were able to inactivate CYP1A inducer(s) in F3 through metabolism. Overall the results suggest that the pattern of CYP1A induction by F3, and by extension, HFO involves complex interactions between the many chemical components in these mixtures. Likely the most important chemicals are the polycyclic aromatic hydrocarbons (PAHs).

A rainbow trout liver cell line, RTL-W1, was used to evaluate the toxic potential of a heavy fuel oil (HFO) HFO 7102, and its fractions, which together with the HFO are referred to as the oil samples. The fractions were F2, F3, F3-1, F3-2 and F4 and had been prepared by low-temperature vacuum distillation by collaborators at Queen's University. For presentation to the cells, HFO 7102 and its fractions were made into High Energy-Chemically Enhanced Water Accommodated Fractions (HE-CEWAFs). The procedure for this involved adding Corexit 9500 to the oil samples, mixing them on a vortex, and letting the phases settle. The HE-CEWAFs were added to RTL-W1 cell cultures, and at various times afterwards cell viability and CYP1A induction were monitored. Cell viability was evaluated with two dyes, Alamar Blue, which monitors energy metabolism, and 5-carboxfluorescein diacetate acetoxymethyl ester (CFDA AM), which measures plasma membrane integrity. With both indicator dyes, Corexit 9500 was cytotoxic but the concentrations eliciting cytotoxicity varied with the cell culture media. In Leibovitz's L-15 with fetal bovine serum (FBS), which was the medium used for studying CYP1A induction, Corexit 9500 was only cytotoxic at concentrations of 0.1 % (v/v) and greater. For the oil samples, F3-2 at 1 mg/ml and F4 at 10 mg/ml, which were the highest testable concentrations for each, no loss of cell viability was observed over 24 h. The other oil samples were cytotoxic only at their highest testable concentrations, which ended being between 1 and 10 mg/ml. CYP1A induction was monitored in RTL-W1 as catalytic activity and as the level of CYP1A (P4501A) protein. The catalytic activity was assayed as 7-ethoxyresorufin o-deethylase (EROD) activity; the CYP1A protein level, by western blotting. The positive control was 2, 3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), which strongly induced both EROD activity and CYP1A protein. Corexit 9500 by itself neither induced EROD activity nor CYP1A protein. All the oil samples induced both EROD activity and CYP1A protein. With both endpoints, the most potent fraction was F3; the least potent, F3-2. As the induction of CYP1A is associated with the development of blue sac disease (BSD) and mortality in early life stages of fish, the results suggest that HFO 7102 and its fractions have the potential to reduce recruitment of young into adult fish populations. CYP1A induction by F3 was studied further, again through EROD activity and western blotting. As the F3 concentration was increased, EROD activity increased but declined at high concentrations, whereas CYP1A protein continued to increase. This suggests the presence of compounds in F3 that at high concentrations inhibit the catalytic activity of CYP1A. When F3 was presented to RTL-W1 cultures together with TCDD, CYP1A protein was induced but not EROD activity. Again this suggests that F3 contains inhibitor(s) of CYP1A as well as inducers. When cultures were exposed to either F3 or TCDD for 24 h and then followed by western blotting for up to 6 days after F3 or TCDD removal, CYP1A levels declined in F3 cultures but not in TCDD cultures. This suggests that RTL-W1 were able to inactivate CYP1A inducer(s) in F3 through metabolism. Overall the results suggest that the pattern of CYP1A induction by F3, and by extension, HFO involves complex interactions between the many chemical components in these mixtures. Likely the most important chemicals are the polycyclic aromatic hydrocarbons (PAHs).

Wuppertal, Universiẗat, Diss., 2005.

No description available.

Chez la majorité des eucaryotes, les petits ARN (miRNA et siRNA) jouent des rôles essentiels au cours du développement, dans les réponses adaptatives aux stress, et dans la maintenance de la stabilité génétique. Les plantes codent quatre enzymes RNaseIII de type DICER-LIKE (DCL). DCL1, produit les miRNAs, tandis que DCL2, DCL3 et DCL4 produisent des siRNAs des tailles diverses. Les plantes codent également des enzymes appelées RNASE-THREE-LIKE (RTL) auxquelles il manque certains domaines spécifiques aux DCLs, et dont la fonction est largement inconnue.Des plantes sur-exprimant RTL1 montrent des défauts morphologiques, et n'accumulent pas les siRNAs produits par DCL2, DCL3 ou DCL4, indiquant que RTL1 est un suppresseur général des voies de siRNA chez les plantes. L’activité de RTL1 nécessite un domaine RNaseIII fonctionnel. RTL1 ne s'exprime naturellement que faiblement dans les racines, mais l'infection virale induite fortement son expression dans les feuilles, ce qui suggère que l’induction de RTL1 est une stratégie générale utilisée par les virus pour contrer la défense antivirale basée sur siRNAs. En accord avec cette hypothèse, les plantes transgéniques sur-exprimant RTL1 sont plus sensibles à l'infection par le TYMV que des plantes de type sauvage, probablement parce que RTL1 empêche la production des siRNAs dirigés contre les RNA viraux. Cependant, les plantes transgéniques sur-exprimant RTL1 ne sont pas plus sensibles à l'infection par le TCV, TVCV ou le CMV, qui codent les suppresseurs de RNA silencing (VSR) plus puissants que le TYMV. En effet, le VSR de TCV inhibe l'activité de RTL1, suggérant que l'induction de l’expression de RTL1 par les virus et l’amortissement de l’activité de RTL1 par leurs VSRs est une double stratégie permettant d’établir une infection avec succès. Des plantes sur-exprimant RTL2 ou des mutants rtl2 ne montrent aucun défaut morphologique, et ne montrent pas de changement majeur du répertoire des petits ARNs endogènes. Toutefois, la sur-expression de RTL2 augmente l’accumulation des petits ARNs exogènes dans des essais d’expression transitoire, et cette activité nécessite un domaine RNaseIII fonctionnel. Il est donc possible que RTL2 clive certains substrats pour faciliter l’action des enzymes DCL. / Small RNAs, including miRNA and siRNA, play essential regulatory roles in genome stability, development and stress responses in most eukaryotes. Plants encode four DICER-LIKE (DCL) RNaseIII enzymes. DCL1 produces miRNAs, while DCL2, DCL3 and DCL4 produce diverse size classes of siRNA. Plants also encode RNASE THREE-LIKE (RTL) enzymes that lack DCL-specific domains and whose function is largely unknown. Arabidopsis plants over-expressing RTL1 exhibit morphological defects and lack all types of small RNAs produced by DCL2, DCL3 and DCL4, indicating that RTL1 is a general suppressor of plant siRNA pathways. RTL1 activity requires a functional RNaseIII domain. RTL1 is naturally expressed only weakly in roots, but virus infection strongly induces its expression in leaves, suggesting that RTL1 induction is a general strategy used by viruses to counteract the siRNA-based plant antiviral defense. Accordingly, transgenic plants over-expressing RTL1 are more sensitive to TYMV infection than wild-type plants, likely because RTL1 prevents the production of antiviral siRNAs. However, TCV, TVCV and CMV, which encode stronger suppressors of RNA silencing (VSR) than TYMV, are insensitive to RTL1 over-expression. Indeed, TCV VSR inhibits RTL1 activity, suggesting that inducing RTL1 expression and dampening RTL1 activity is a dual strategy used by viruses to establish a successful infection. Plants over-expressing RTL2 and rtl2 mutants do not exhibit morphological defects and do not show major changes in the endogenous small RNA repertoire. However, RTL2 over-expression enhances the accumulation of exogenous siRNAs in transient assays, and this activity requires a functional RNaseIII domain. Therefore, it is possible that plant RTL2 processes certain substrates to facilitate the action of DCL enzymes.