Neural network based off-line handwritten text recognition system

This dissertation introduces a new system for handwritten text recognition based on an improved neural network design. Most of the existing neural networks treat mean square error function as the standard error function. The system as proposed in this dissertation utilizes the mean quartic error function, where the third and fourth derivatives are non-zero. Consequently, many improvements on the training methods were achieved. The training results are carefully assessed before and after the update. To evaluate the performance of a training system, there are three essential factors to be considered, and they are from high to low importance priority: (1) error rate on testing set, (2) processing time needed to recognize a segmented character and (3) the total training time and subsequently the total testing time. It is observed that bounded training methods accelerate the training process, while semi-third order training methods, next-minimal training methods, and preprocessing operations reduce the error rate on the testing set. Empirical observations suggest that two combinations of training methods are needed for different case character recognition. Since character segmentation is required for word and sentence recognition, this dissertation provides also an effective rule-based segmentation method, which is different from the conventional adaptive segmentation methods. Dictionary-based correction is utilized to correct mistakes resulting from the recognition and segmentation phases. The integration of the segmentation methods with the handwritten character recognition algorithm yielded an accuracy of 92% for lower case characters and 97% for upper case characters. In the testing phase, the database consists of 20,000 handwritten characters, with 10,000 for each case. The testing phase on the recognition 10,000 handwritten characters required 8.5 seconds in processing time.

Predictors of developmental gains in children enrolled in an early intervention setting

The purpose of this research was to determine whether initial developmental delay, site of intervention, frequency of intervention, age of the child, socio-economic status (SES), gender and ethnicity significantly predict developmental gains in a group of children enrolled in an early intervention setting. The records of 134 children enrolled in an inner-city program in Miami, Florida were reviewed for inclusion in this study. ^ Demographic variables, site placement and treatment frequencies were collected during a retrospective chart review. Level of delay was expressed using the developmental quotient and developmental gain was calculated using the mean gain on age equivalent scores or developmental tests. A multiple regression analysis was performed to determine which of the above variables significantly predicted developmental gains. Multivariate analysis compared developmental gains for all the developmental domains based on intervention site (center versus home-based) while controlling for developmental delay. ^ Children made greater developmental gains if they had higher developmental quotients and if they were younger at the time services were initiated. Frequency of intervention significantly improved developmental outcomes in children attending center-based programs. Children attending center-based programs also made significantly greater gains in gross motor skills compared to children attending home-based programs. ^ These findings emphasize the importance of early screening and referral of children with developmental delay and adjusting intervention for the child's developmental quotient. Children should receive intense treatment to maximize results. Decisions regarding program placement should be individualized according to the child's unique developmental pattern. Policy and program decisions affecting the curriculum of a child in early intervention need to reflect these multivariate considerations. ^

Diversity of Hymenoptera, cultivated plants and management practices in home garden agroecosystems, Kyrgyz Republic

Pollination-dependent fruit trees grown in home gardens play an important role in the agricultural based economy of Central Asian countries, yet little is known about the status of pollinator communities, the cultivated plant composition or the factors that influence management practices in Kyrgyz home garden agroecosystems. As agricultural systems are human created and managed, a logical approach to their study blends anthropological and ecological methods, an ethnoecological approach. Over three years, I investigated how species richness and abundance of Hymenoptera, cultivated plants, and home garden management were related using quantitative and qualitative methods in the Issyk-kul Man and Biosphere reserve. Structured surveys were undertaken with heads of households using a random sample stratified by village. Gardens were then mapped with participation of household members to inventory edible species in gardens, most of which are pollinator-dependent, and to compare home garden diversity as reported by respondents during interviews. Apple diversity was studied to the variety level to understand respondents’ classification system in the context of in situ agrobiodiversity conservation. Household members identified 52 edible plant species when mapping the garden, compared with 32 reported when interviewed. The proportion of plant species received from others through exchange and the number of plots cultivated significantly explained the variation in edible plant diversity among gardens. Insects were sampled in gardens and orchards to determine potential pollinator community composition and the effect of different management practices on Hymenoptera richness and abundance. I collected 756 Hymenoptera individuals (56 bee; 12 wasp species); 12 species were new records for Kyrgyzstan or within Kyrgyzstan. Economic pressures to intensify cultivation could impact management practices that currently promote diversity. A home garden development initiative was undertaken to study management practice improvement. Participants in the initiative had higher adoption rates than controls of management practices that improve long-term yield, ecological sustainability and stability of home gardens. Home gardens, as currently managed, support abundant and diverse pollinator communities and have high cultivated plant diversity with few differences in community composition between garden management types.

Analysis of Heterokaryon Formation in Fusarium oxysporum f. sp. cubense.

Fusarium oxysporum forma specialis cubense is a soilborne phytopathogen that infects banana. The true evolutionary identity of this so called species, Fusarium oxysporum, is still unknown. Many techniques have been applied in order to gain insight for the observed genetic diversity of this species. The current classification system is based on vegetative compatibility groups (VCG's). Vegetative compatibility is a self non-self recognition system in which only those belonging to a VCG can form stable heterokaryons, cells containing two distinct nuclei. Heterokaryons in turn, are formed from hypha! anastomosis, the fusion of two hyphae. Furthermore, subsequent to heterokaryon formation potential mechanisms exist which may generate genetic variability. One is through viral transfer upon hyphal anastomosis. The other mechanism is a form of mitotic recombination referred to as the parasexual cycle. Very little research has been performed to directly obser.ve the cellular events; hypha! anastomosis, heterokaryon formation, and the parasexual cycle in Fusarium oxysporum f. sp. cubense. The purpose of this research was to design and use methods which would allow for the detection of hypha! anastomosis and heterokaryon formation, as well as any characteristics surrounding this event, within and between VCG's in Foe. First, some general growth properties were recorded: the number of nuclei per hypha, the size ofthe hyphal tip cell, the size of the cell adjacent to the hypha! tip (pre-tip) cell, and the number of cells to the first branch point. Second, four methods were designed in order to assay hyphal anastomosis and heterokaryon formation: 1) pairings on membrane: phase or brightfield microscopy, 2) pairings on membrane: fluorescence microscopy, 3) spore crosses: fluorescence microscopy, and 4) double picks in fractionated MMA. All of these methods were promtsmg.

The Role of GIS in DRR: Classifying and Improving GIS Capabilities in Small Towns and Rural Areas

Optimizing GIS capability does not always require that the municipality obtain cutting edge professionals and resources. This paper offers a disaster risk reduction (DRR) design methodology for small towns and rural areas that employs a multi-variable classification system, enabling customization for effective DRR. Determining appropriate GIS capacity requires that a community first be evaluated in order to identify its disaster risk reduction/disaster management (DRR/DM) requirements. These requirements are then considered in conjunction with the municipality's resources to establish the desired capability. Qualification levels for major aspects of GIS capability with respect to DRR/DM are provided along with descriptions of each level and suggested procedures for advancement to the next level. It should be noted that a municipality can be classified at a different level with respect to different variables. Needs vary according to the community, thus attainment of a uniform capability level may not be necessary.

Diversity of Hymenoptera, Cultivated Plants and Management Practices in Home Garden Agroecosystems, Kyrgyz Republic

Pollination-dependent fruit trees grown in home gardens play an important role in the agricultural based economy of Central Asian countries, yet little is known about the status of pollinator communities, the cultivated plant composition or the factors that influence management practices in Kyrgyz home garden agroecosystems. As agricultural systems are human created and managed, a logical approach to their study blends anthropological and ecological methods, an ethnoecological approach. Over three years, I investigated how species richness and abundance of Hymenoptera, cultivated plants, and home garden management were related using quantitative and qualitative methods in the Issyk-kul Man and Biosphere reserve. Structured surveys were undertaken with heads of households using a random sample stratified by village. Gardens were then mapped with participation of household members to inventory edible species in gardens, most of which are pollinator-dependent, and to compare home garden diversity as reported by respondents during interviews. Apple diversity was studied to the variety level to understand respondents’ classification system in the context of in situ agrobiodiversity conservation. Household members identified 52 edible plant species when mapping the garden, compared with 32 reported when interviewed. The proportion of plant species received from others through exchange and the number of plots cultivated significantly explained the variation in edible plant diversity among gardens. Insects were sampled in gardens and orchards to determine potential pollinator community composition and the effect of different management practices on Hymenoptera richness and abundance. I collected 756 Hymenoptera individuals (56 bee; 12 wasp species); 12 species were new records for Kyrgyzstan or within Kyrgyzstan. Economic pressures to intensify cultivation could impact management practices that currently promote diversity. A home garden development initiative was undertaken to study management practice improvement. Participants in the initiative had higher adoption rates than controls of management practices that improve long-term yield, ecological sustainability and stability of home gardens. Home gardens, as currently managed, support abundant and diverse pollinator communities and have high cultivated plant diversity with few differences in community composition between garden management types.

Neuromuscular Changes in Older Adults during the Lateral Step Task

Older adults may have trouble when performing activities of daily living due to decrease in physical strength and degradation of neuromotor and musculoskeletal function. Motor activation patterns during Lateral Step Down and Step Up from 4-inch and 8-inch step heights was assessed in younger (n=8, 24.4 years) and older adults (n=8, 58.9 years) using joint angle kinematics and electromyography of lower extremity muscles. Ground reaction forces were used to ascertain the loading, stabilization and unloading phases of the tasks. Older adults had an altered muscle activation sequence and significantly longer muscle bursts during loading for the tibialis anterior, gastrocnemius, vastus medialis, bicep femoris, gluteus medius and gluteus maximus muscles of the stationary leg. They also demonstrated a significantly larger swing time (579.1 ms vs. 444.8 ms) during the step down task for the moving leg. The novel data suggests presence of age-related differences in motor coordination during lateral stepping.