• Journal of Multimedia Information System
• /
• v.9 no.3
• /
• pp.183-190
• /
• 2022

Image-based 3D object detection is one of the important and difficult problems in autonomous driving and robotics, and aims to find and represent the location, dimension and orientation of the object of interest. It generates three dimensional (3D) bounding boxes with only 2D images obtained from cameras, so there is no need for devices that provide accurate depth information such as LiDAR or Radar. Image-based methods can be divided into three main categories: monocular, stereo, and multi-view 3D object detection. In this paper, we investigate the recent state-of-the-art models of the above three categories. In the multi-view 3D object detection, which appeared together with the release of the new benchmark datasets, NuScenes and Waymo, we discuss the differences from the existing monocular and stereo methods. Also, we analyze their performance and discuss the advantages and disadvantages of them. Finally, we conclude the remaining challenges and a future direction in this field.

• Proceedings of the Optical Society of Korea Conference
• /
• 2005.07a
• /
• pp.204-205
• /
• 2005

Studies on display has been requested some major changes due to the high growth of the handheld terminal market. Therefore, the self emitting OLED(Organic Light Emitting Diode) has been interested as a next generation flat plane display because of its preeminent characteristics such as quick response characteristics, higher performance viewing angle, low power consumption, and panel floating. However, a trend of the display market is moving to three dimensional image processing instead of two dimensional flat display and various researches on display using hologram makes up for the difficulty in three dimensional display using typical flat display. In this study the Lenticular Screen Printing method is presented so that it can be applicable to organic semiconductor display devices and makes possible three dimensional display using flat display for complement the drawback of inorganic semiconductor.

• Korean Journal of Acupuncture
• /
• v.25 no.4
• /
• pp.167-174
• /
• 2008

Objectives : Acupuncture points and meridians have been usually depicted as a two dimensional drawing and verbal description. Recently, imaging and three-dimensional image processing technologies have been introduced into medical fields such as anatomy and virtual operation, for the purpose of enhanced efficiency in research and education. This study attempted an image modelling of the meridian and acupoint in the upper limb region. Methods : A vector image model of an arm was produced and medical information on the meridian and acupoint of the arm region was incorporated. Results : A 3D modelling of the acupuncture meridian and acupoint in the upper limb region was produced along with a user console to control the presentation of related information and to facilitate visualization of the 3D model images. Conclusions : A 3D modelling of the acupuncture meridian and acupoint will be an efficient platform for an education and research.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1997.11a
• /
• pp.79-85
• /
• 1997

Virtual holography is a methodology of synthesizing apparent three-dimensional images from two-dimensional photographs. Since the input is photographic images of real objects, the degree of realism exceeds that offered by any computer-aided design software. The three-dimensional appearance is given in real-time by images from arbitrary viewing directions. If infinitely many photographs were taken and pasted together, virtual holography would have been trivial. But, the (infinite) storage requirement would prohibit such an attempt.

• Journal of Broadcast Engineering
• /
• v.23 no.2
• /
• pp.235-245
• /
• 2018

Recently panoramic image overcomes camera limited viewing angle and offers wide viewing angle by stitching plenty of images. In this paper, we propose pre-processing and post-processing algorithm which makes speed and accuracy improvements when making panoramic images. In pre-processing, we can get camera sensor information and use three-dimensional rotation angle to find RoI(Region of Interest) image. Finding RoI images can reduce time when extracting feature point. In post-processing, we propose weighted minimal error boundary cut blending algorithm to improve accuracy. This paper explains our algorithm and shows experimental results comparing with existing algorithms.

• Journal of Biomedical Engineering Research
• /
• v.24 no.4
• /
• pp.247-257
• /
• 2003

This paper proposes a three-dimensional (3D) segmentation algorithm for extracting a diagnostic object from ultrasound images by using a LoG operator In the proposed algorithm, 2D cutting planes are first obtained by the equiangular revolution of a cross sectional Plane on a reference axis for a 3D volume data. In each 2D ultrasound image. a region of interest (ROI) box that is included tightly in a diagnostic object of interest is set. Inside the ROI box, a LoG operator, where the value of $\sigma$ is adaptively selected by the distance between reference points and the variance of the 2D image, extracts edges in the 2D image. In Post processing. regions of the edge image are found out by region filling, small regions in the region filled image are removed. and the contour image of the object is obtained by morphological opening finally. a 3D volume of the diagnostic object is rendered from the set of contour images obtained by post-processing. Experimental results for a tumor and gall bladder volume data show that the proposed method yields on average two times reduction in error rate over Krivanek's method when the results obtained manually are used as a reference data.

• The Journal of the Acoustical Society of Korea
• /
• v.19 no.3
• /
• pp.20-25
• /
• 2000

This paper proposes the method of surface classification and threshold value selection for surface classification of the three-dimensional object recognition. The processings of three-dimensional image processing system consist of three steps, i.e, acquisition of range data, feature extraction and matching process. This paper proposes the method of shape feature extraction from the acquired rage data in the entire three-dimensional image processing system. In order to achieve these goals, firstly, this article proposes the surface classification method by using the distribution characteristics of sign value from range values. Also pre-existing method which uses the K-curvature and K-curvature has limitation in the practical threshold value selection. To overcome this, this article proposes the selection of threshold value for surface classification. Finally, the effectiveness of this article is demonstrated by the several experiments.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.11
• /
• pp.1664-1671
• /
• 2004

It is necessary to have a model that describes the feature of the knee Joint with a sufficient accuracy. Koreans, however, do not have their own knee joint model to be used in the total knee replacement arthroplasty. They have to use European or American models which do not match Koreans. Three-dimensional visualization techniques are found to be useful in a wide range of medical applications. Three-dimensional imaging studies such as CT(computed tomography) and MRI(magnetic resonance image) provide the primary source of patient-specific data. Three-dimensional knee joint models were constructed by image processing of the CT data of 10 subjects. Using the constructed model, the dimensions of Korean knee joint were measured. And this study proposed a three-dimensional model and data, which can be helpful to develop Korean knee implants and to analyze knee joint movements.

• Applied Microscopy
• /
• v.36 no.3
• /
• pp.217-226
• /
• 2006

Image processing by ultra high voltage electron microscopy (UHVEM) and tomography has offered major contributions to research in the field of cellular ultrastructure. Furthermore, such advancements also have enabled the improved analysis of three-dimensional cellular structures in botany. In the present study. using UHVEM and tomography, we attempted to reconstruct the three-dimensional images of plastid inclusions that probably differentiate during photosynthesis. The foliar tissues were studied Primarily with the TEM and further examined with UHVEM. The spatial relationship between tubular elements and the thylakoidal membrane and/or starch grains within plastids mainly have been investigated in CAM-performing Sedum as well as in $C_4$ Salsola species. The inclusion bodies were found to occur only in early development in the former, while they were found only in mesophyll cells in the latter. The specimens were tilted every two degrees to obtain two-dimensional images with UHVEM and subsequently comparison has been made between the two types. Digital image processing was performed on the elements of the inclusion body using tilting, tomography, and IMOD program to generate and reconstruct three-dimensional images on the cellular level. In Sedum plastids, the inclusion bodies consisted of tubular elements exhibiting about 20 nm distance between elements. However, in Salsola, plastid inclusion bodies demonstrated quite different element structure, displaying pattern, and origin relative to those of the Sedum. The inclusion bodies had an integrative relationship with the starch grains in both species.

• Journal of Biosystems Engineering
• /
• v.14 no.2
• /
• pp.104-114
• /
• 1989

The most important subject in developing agricultural robots for fruit harvesting is to detect accurately the location of a fruit about the given coordinate system. This study was carried out to develop an image processing algorithm which enables finding out the three dimensional locations of a fruit. The digital image processing device consisted of an optosensor (Closed-circuit TV camera), image processing interface board (Digitizer) and microcomputer (IBM PC/AT). A stereo-image processing method using the two cameras attached to the manipulator was evaluated through experiment with apples. The accuracy and quickness of detecting the location of apples by this method was not satisfactory. The maximum errors of the detected locations by the stereo-image processing method in x-, Y-, and z- directions were 3, 4 and 4 cm, respectively. The maximum time required to get the rectangular coordinate data of a fruit was about 2 minutes.