• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.18 no.9
• /
• pp.81-87
• /
• 2019

The automatic tool changer of a machining center consists of a tool magazine and a cam box, and the core components of the cam box are a roller gear cam and a turret. Recently, the roller gear cam of a cam box has been replaced by a ball gear cam. In this study, the design and machining method of ball gear cam for an automatic tool changer was studied. Additionally, an algorithm for a 4-axis post processing method was established from an instrumental formula by designing a ball gear cam, thus preventing machining at the bottom of ball end mill and enabling the ball on the turret to be driven at the entrance and exit of a curve without collision due to machining errors. In conclusion, machining using only the 4-axis method including the C-axis on a BC -Type 5-axis machine produced the desired ball gear cam.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.16 no.4
• /
• pp.145-151
• /
• 2021

In this paper, we propose a novel visualization technique to explain the predictions of deep neural networks. We use knowledge distillation (KD) to identify the interior of a black-box model for which we know only inputs and outputs. The information of the black box model will be transferred to a white box model that we aim to create through the KD. The white box model will learn the representation of the black-box model. Second, the white-box model generates attention maps for each of its layers using Grad-CAM. Then we combine the attention maps of different layers using the pixel-wise summation to generate a final saliency map that contains information from all layers of the model. The experiments show that the proposed technique found important layers and explained which part of the input is important. Saliency maps generated by the proposed technique performed better than those of Grad-CAM in deletion game.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.4 no.3
• /
• pp.121-125
• /
• 2003

A switch box is metal box for electric wire pipes in building. In manufacturing this box, the first press is used to mold and the second to fifth presses are used to process four-sides of a box. So four presses, four molds and four workers are needed. This leads to raise the manufacturing cost including machine lease rent, molding fee, and wages. To make matters worse, the roaring sound, vibration, and dangers of safety accidents make employees avoid from working that job. To solve this problem, a new Handling Mechanism and a cam die which combined four-die-involving processes into one die process were developed. That results job automation of the Job with one press and one die. The job avoidance problem is solved and manufacturing cost is reduced as well.

• Smart Media Journal
• /
• v.12 no.9
• /
• pp.45-59
• /
• 2023

Weakly supervised object localization (WSOL) is a task of localizing an object in an image using only image-level labels. Previous studies have followed the conventional class activation mapping (CAM) pipeline. However, we reveal the current CAM approach suffers from problems which cause original CAM could not capture the complete defects features. This work utilizes a convolutional neural network (CNN) pretrained on image-level labels to generate class activation maps in a multi-scale manner to highlight discriminative regions. Additionally, a vision transformer (ViT) pretrained was treated to produce multi-head attention maps as an auxiliary detector. By integrating the CNN-based CAMs and attention maps, our approach localizes defective regions without requiring bounding box or pixel-level supervision during training. We evaluate our approach on a dataset of apple images with only image-level labels of defect categories. Experiments demonstrate our proposed method aligns with several Object Detection models performance, hold a promise for improving localization.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.9
• /
• pp.349-358
• /
• 2021

Recently, a method of finding the attention area or localization area for an object of an image using CAM (Class Activation Map)[1] has been variously carried out as a study of WSOL (Weakly Supervised Object Localization). The attention area extraction from the object heat map using CAM has a disadvantage in that it cannot find the entire area of the object by focusing mainly on the part where the features are most concentrated in the object. To improve this, using CAM and Selective Search[6] together, we first expand the attention area in the heat map, and a Gaussian smoothing is applied to the extended area to generate retraining data. Finally we train the data to expand the attention area of the objects. The proposed method requires retraining only once, and the search time to find an localization area is greatly reduced since the selective search is not needed in this stage. Through the experiment, the attention area was expanded from the existing CAM heat maps, and in the calculation of IOU (Intersection of Union) with the ground truth for the bounding box of the expanded attention area, about 58% was improved compared to the existing CAM.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.19 no.12
• /
• pp.98-104
• /
• 2020

In this work, a study on the 5-axis machining of ball gear cam is conducted which is a continuation of reference [1]. The ball gear cam used in this study delivers motion in conjunction with the ball supported by the turret. Therefore, it requires carbonizing heat treatment and is usually completed using a 4-axis machining with a carbide ball end mill. If the nose part of the ball end mill is not allowed to participate in the machining, then CBN tools without the nose part can be used. However, machining of certain shapes can be carried out only by contacting the ball in some of the areas on either side which can improve the surface of the machining. This requires a 5-axis machining in order to maintain a constant angle for the processing path. Therefore, in this work, the 5-axis machining method is studied in order to maintain the direction of the cutter axis at a constant angle with the tangent direction of the curve-ball gear cam. Furthermore, the 5-axis machining program for the ball gear cam was developed and the machining experiment was completed and verified.

• Journal of the Korean Solar Energy Society
• /
• v.31 no.1
• /
• pp.44-50
• /
• 2011

The solar radiation, inner temperature of the tilted box and the amount of condensed water were measured to find the most effective purified water production system. The inner temperature of the tilted box was related with solar radiation that was affected by sunshine and cloud. With analyzing those relationships of the solar radiation, inner box temperature, the angle of tilted boxes and amounts of water condensation optimal collecting system area could be calculated.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2012.05a
• /
• pp.269-269
• /
• 2012

• BMB Reports
• /
• v.41 no.11
• /
• pp.771-777
• /
• 2008

Calmodulin (CaM) proteins, members of the EF-hand family of $Ca^{2+}$-binding proteins, represent important relays in plant calcium signals. Here, OsCam1-1 was isolated by PCR amplification from the rice genome. The gene contains an ORF of 450 base pairs with a single intron at the same position found in other plant Cam genes. A promoter region with a TATA box at position-26 was predicted and fused to a gus reporter gene, and this construct was used to produce transgenic rice by Agrobacterium-mediated transformation. GUS activity was observed in all organs examined and throughout tissues in cross-sections, but activity was strongest in the vascular bundles of leaves and the vascular cylinders of roots. To examine the properties of OsCaM1-1, the encoding cDNA was expressed in Escherichia coli. The electrophoretic mobility shift when incubated with $Ca^{2+}$ indicates that recombinant OsCaM1-1 is a functional $Ca^{2+}$-binding protein. In addition, OsCaM1-1 bound the CaMKII target peptide confirming its likely functionality as a calmodulin.

• International Journal of CAD/CAM
• /
• v.3 no.1_2
• /
• pp.85-95
• /
• 2003

During the design of dies and molds, the cavity of the object is obtained by subtracting it from a surrounding rectangular block. This box is subsequently split into two halves by the parting surface. Similar problems also occur in some RP processes such as LOM, SGC, SLS and 3DP where the machine produces a block inside which the prototype is buried. Determining the orientation of the object inside the box and the corresponding parting surface taking appropriate constraints into account have been addressed by several researchers. However, given the parting surface, the problem of splitting the box development of a software package called OptiLOM (now a module of an RP software Magics 8.0), the authors realized non-triviality of this problem since the loop can spread over as many as 5 faces of the box. In this paper, the authors have tried to bring out the importance of this problem and have presented their algorithm to solve it.