|Year : 2018 | Volume
| Issue : 1 | Page : 106-109
Use and validation of mirrorless digital single light reflex camera for recording of vitreoretinal surgeries in high definition
Sumeet Khanduja1, Raju Sampangi2, BC Hemlatha3, Satvir Singh4, Ashish Lall5
1 Department of Ophthalmology, Kalpana Chawla Government Medical College, Karnal, Haryana, India
2 Department of Ophthalmology, Netraspandana Eye Hospital, Bengaluru, Karnataka, India
3 Department of Ophthalmology, Minto Regional Institute of Ophthalmology, Bengaluru, Karnataka, India
4 Department of Ophthalmology, Regional Institute of Ophthalmology, Pt. B. D. Sharma Post Graduate Institute of Medical Sciences, Rohtak, Haryana, India
5 Lall Eye and Skin Care Center, Gurgaon, Haryana, India
|Date of Submission||18-Jun-2017|
|Date of Acceptance||05-Oct-2017|
|Date of Web Publication||28-Dec-2017|
Dr. Sumeet Khanduja
Department of Ophthalmology, Kalpana Chawla Government Medical College, Karnal - 132 001, Haryana
Source of Support: None, Conflict of Interest: None
Purpose: The purpose of this study is to describe the use of commercial digital single light reflex (DSLR) for vitreoretinal surgery recording and compare it to standard 3-chip charged coupling device (CCD) camera. Methods: Simultaneous recording was done using Sony A7s2 camera and Sony high-definition 3-chip camera attached to each side of the microscope. The videos recorded from both the camera systems were edited and sequences of similar time frames were selected. Three sequences that selected for evaluation were (a) anterior segment surgery, (b) surgery under direct viewing system, and (c) surgery under indirect wide-angle viewing system. The videos of each sequence were evaluated and rated on a scale of 0-10 for color, contrast, and overall quality Results: Most results were rated either 8/10 or 9/10 for both the cameras. A noninferiority analysis by comparing mean scores of DSLR camera versus CCD camera was performed and P values were obtained. The mean scores of the two cameras were comparable for each other on all parameters assessed in the different videos except of color and contrast in posterior pole view and color on wide-angle view, which were rated significantly higher (better) in DSLR camera. Conclusion: Commercial DSLRs are an affordable low-cost alternative for vitreoretinal surgery recording and may be used for documentation and teaching.
Keywords: Medical education, medical innovation, ophthalmic surgery recording
|How to cite this article:|
Khanduja S, Sampangi R, Hemlatha B C, Singh S, Lall A. Use and validation of mirrorless digital single light reflex camera for recording of vitreoretinal surgeries in high definition. Indian J Ophthalmol 2018;66:106-9
|How to cite this URL:|
Khanduja S, Sampangi R, Hemlatha B C, Singh S, Lall A. Use and validation of mirrorless digital single light reflex camera for recording of vitreoretinal surgeries in high definition. Indian J Ophthalmol [serial online] 2018 [cited 2019 Sep 19];66:106-9. Available from: http://www.ijo.in/text.asp?2018/66/1/106/221788
Video-assisted skill transfer is very effective way to train residents during ophthalmic training. Numerous companies such as Sony, Ikegami, and Optronics manufacture high quality recording systems. However, cost constraints come in the way of procuring these quality high definition (HD) recording systems. To overcome this, we explored various cost-effective alternatives such as single chip charged coupling device (CCD) and webcam-based cameras. Although these cameras give the acceptable quality video for recording anterior segment surgery, we observed that they are not suited for recording vitreoretinal surgeries. This is because of the difference in illumination sources used during surgery. A bright light emitting diode light from the microscope is used for anterior segment surgeries while retinal surgeries are essentially performed in low light using the light from a fine endoilluminator probe. Hence, vitreoretinal surgery recording needs a camera sensor with three qualities: (1) high quantum efficiency so that it can capture videos in low light, (2) high signal to noise ratio so that these images do not have grain, and (3) broad dynamic range so that both well-illuminated and dark areas are visualized equally. Only the high-end 3-chip CCD medical cameras (priced nearly USD 21,000) fulfilled these exacting requirements till now. Recently, high-end digital single light reflex (DSLR's) cameras have been developed, which fulfill these criteria and are reasonably priced (approximately USD 2500) compared to HD medical cameras. These cameras are used extensively for high-resolution videography and low light astrophotography. We herein describe the use of these DSLRs and as an affordable alternative to the CCD cameras for ophthalmic surgery, vitreoretinal surgical recording in particular and compare the recording quality to the current gold standard, i.e., the HD 3-chip CCD cameras.
| Methods|| |
Recording technique: Camera – Why Sony A7S2?
Among the DSLRs, various options were considered from Canon, Nikon, and Panasonic we finally zeroed in on Sony A7S2. It is a mirrorless, full frame DSLR camera as it had the best internet reviews on low-light performance ,, which is required for recording vitreoretinal surgeries. It has the following specifications (a) ISO up to 400000 (b) record videos in full HD and 4K resolution (c) internal 4K recording to an SD card (no additional external recorder required for recording 4K videos [d]) Simultaneous 4K and HD dual recording.
A Sony A7S 2 mirrorless DSLR is latched to a DSLR adapter (3D Medisys India) for Sony E-mount cameras. The adapter is a fabrication fitted with an array of lenses to focus the light rays diverted from the beam splitter onto the camera sensor. The adapter is latched to the camera body in the same way as any lens is done. The other end on the adapter is inserted in the beam splitter tube and is secured in place with a tightening ring. A labeled diagram of the DSLR camera latching assembly is shown in [Figure 1] Then, the camera is switched on and is put in the APS-C sensor mode. The area of interest is focused, and the same is confirmed on the camera screen. The adapter has a provision for fine focusing which may be used if necessary. Slight vignetting (dark borders at the edges of the frame) is overcome by digitally magnifying the image. Recording is usually done in the manual mode as it gives us control over the brightness of the recording by varying the ISO. In our experience, we have seen that anterior segment surgery recording is best done at an ISO of 200–400 and vitreoretinal surgery recording under a direct viewing system such as irrigating contact lens is best achieved at an ISO of 400–600 and an ISO of 1600–2500 does best for recording under wide-angle viewing systems. The camera gives a facility to customize the settings by the fixing the upper and lower range and the amount of exposure (brightness) of any recording.
|Figure 1: Labeled photograph showing various components of the digital single light reflex recording assembly|
Click here to view
Simultaneous recording of vitreoretinal surgery was done using two camera systems. The surgery was performed on Zeiss Opmi 140 microscope with DORC Associate posterior vitrectomy system. The simultaneous recording assembly comprised of a Sony HD 3-chip CCD camera (Model no. PMW-10MDC) with Sony medical recorder (coded camera 1) on one side of the beam splitter and the Sony A7s2 camera (coded camera 2) on the other side. The videos recorded from both the camera systems were edited for removing redundant segments and sequences of similar time frames were selected. No editing was done for the video size, colors, contrast, dynamic range adjustment, blur, flicker, or motion blur. Three sequences that selected for evaluation were (a) anterior segment surgery, (b) surgery under direct viewing system, and (c) surgery under indirect wide-angle viewing system the corresponding video clips were played simultaneously on a laptop having a diagonal screen width of 15 inches against a gray background. Twenty evaluators (practicing vitreoretinal surgeons with postfellowship experience between 5 and 10 years) were seated at a distance of 1.5 meters (approximately 8H where H = height of an individual video) from the screen. Three videos of each sequence were evaluated and rated on a scale of 0-10 [Figure 2] for color, contrast and overall video quality. The evaluators were blinded to the type of camera used for a particular video.
Data were entered into Microsoft Excel and mean scores were calculated for scores given to three different videos of each parameter for a specific camera. Further, these data were analyzed using Statistical Package for Social Sciences Version 21. The mean scores of camera 1 were compared with mean scores of camera 2 using Pearson's correlation coefficient and noninferiority was established individually for various parameters with a nonsignificant P value (>0.05) for all results of correlation tests.
| Results|| |
A total of 60 score values for each parameter (color, contrast, and quality) were obtained for each camera – three different eye videos per camera rated by twenty observers. Most results were rated either 8/10 or 9/10 for both the cameras. Frequency percentage of these scores out of 60 is reported in [Table 1].
A noninferiority analysis by comparing mean scores of camera 1 versus camera 2 was performed and P values were obtained. These have been described in [Table 2]. The mean scores of camera 1 and camera 2 were comparable for each other on all parameters assessed in the different videos except of color and contrast in posterior pole view and color on wide angle view which were rated significantly higher in camera 2. Thus, it was established that camera 2 was noninferior when compared with camera 1. The sample photographs taken as a snapshot from the videos from the two cameras is displayed in [Figure 3],[Figure 4],[Figure 5].
|Table 2: P values indicating difference in mean scores for various parameters for anterior segment, posterior pole, and wide angle view obtained from the two cameras|
Click here to view
|Figure 3: Anterior segment snapshot from Sony 3-chip charged coupling device camera (above) and Sony A7s2 (below)|
Click here to view
|Figure 4: Posterior segment posterior pole snapshot from Sony 3-chip charged coupling device camera (above) and Sony A7s2 (below)|
Click here to view
|Figure 5: Posterior segment wide angle snapshot under air from Sony 3-chip charged coupling device camera (above) and Sony A7s2 (below)|
Click here to view
| Discussion|| |
The concept of utilizing commercial cameras for filming eye surgery is not new. It has been described previously for documenting oculofacial and anterior segment surgery. To the best of our knowledge, this is the first report of use and validation of a commercial camera for recording vitreoretinal surgery.
The difference between a medical camera and a commercial camera lies in the sensor technology. Medical cameras use CCD sensors while the commercial cameras rely on complementary metal oxide semiconductor (CMOS) sensors. The CCD sensors have high quantum efficiency (up to 90%) and were widely used for high-quality image acquisition such as broadcasting, space shuttle photography, and medical recording. Recently, interest in CMOS technology was renewed with the expansion of low cost and less exacting digital imaging technology in the form of mobile phone cameras. The technological refinement of CMOS sensors has leapfrogged in the last decade to the point that today all DSLRs and professional movie cameras employ CMOS sensors instead of CCD sensors. Low-light image capture of these sensors is enhanced by the use of a technology called backside illumination, which increases the quantum efficiency of the sensor plate from 60% to 90%. Sony was the first company to introduce backlit sensors for commercial use. The called them as the EXMOR R sensors, and it was this technology that was integrated into our DSLR. In our personal experience, we have found the Sony A7S 2 to perform better than other DSLRs from established companies. One limitation of this technology is that there is no provision for overlaying parameter display from the operating microscope or the operating machine which may be a limitation for using the camera for live surgery at the conference and meets.
| Conclusion|| |
We believe that commercial DSLRs are an affordable low-cost alternative for vitreoretinal surgery recording and may be used for skill demonstration and teaching
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Morris T. Modifying DSLR cameras for astrophotography. J Br Astron Assoc 2012;122:162-6.
Lin LK. Surgical video recording with a modified Gopro Hero 4 camera. Clin Ophthalmol 2016;10:117-9.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2]